XLNetΒΆ
OverviewΒΆ
The XLNet model was proposed in XLNet: Generalized Autoregressive Pretraining for Language Understanding by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le. XLnet is an extension of the Transformer-XL model pre-trained using an autoregressive method to learn bidirectional contexts by maximizing the expected likelihood over all permutations of the input sequence factorization order.
The abstract from the paper is the following:
With the capability of modeling bidirectional contexts, denoising autoencoding based pretraining like BERT achieves better performance than pretraining approaches based on autoregressive language modeling. However, relying on corrupting the input with masks, BERT neglects dependency between the masked positions and suffers from a pretrain-finetune discrepancy. In light of these pros and cons, we propose XLNet, a generalized autoregressive pretraining method that (1) enables learning bidirectional contexts by maximizing the expected likelihood over all permutations of the factorization order and (2) overcomes the limitations of BERT thanks to its autoregressive formulation. Furthermore, XLNet integrates ideas from Transformer-XL, the state-of-the-art autoregressive model, into pretraining. Empirically, under comparable experiment settings, XLNet outperforms BERT on 20 tasks, often by a large margin, including question answering, natural language inference, sentiment analysis, and document ranking.
Tips:
The specific attention pattern can be controlled at training and test time using the perm_mask input.
Due to the difficulty of training a fully auto-regressive model over various factorization order, XLNet is pretrained using only a sub-set of the output tokens as target which are selected with the target_mapping input.
To use XLNet for sequential decoding (i.e. not in fully bi-directional setting), use the perm_mask and target_mapping inputs to control the attention span and outputs (see examples in examples/text-generation/run_generation.py)
XLNet is one of the few models that has no sequence length limit.
The original code can be found here.
XLNetConfigΒΆ
-
class
transformers.
XLNetConfig
(vocab_size=32000, d_model=1024, n_layer=24, n_head=16, d_inner=4096, ff_activation='gelu', untie_r=True, attn_type='bi', initializer_range=0.02, layer_norm_eps=1e-12, dropout=0.1, mem_len=None, reuse_len=None, bi_data=False, clamp_len=- 1, same_length=False, summary_type='last', summary_use_proj=True, summary_activation='tanh', summary_last_dropout=0.1, start_n_top=5, end_n_top=5, pad_token_id=5, bos_token_id=1, eos_token_id=2, **kwargs)[source]ΒΆ This is the configuration class to store the configuration of a
XLNetModel
. It is used to instantiate an XLNet 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 xlnet-large-cased architecture.Configuration objects inherit from
PretrainedConfig
and can be used to control the model outputs. Read the documentation fromPretrainedConfig
for more information.- Parameters
vocab_size (
int
, optional, defaults to 32000) β Vocabulary size of the XLNet model. Defines the different tokens that can be represented by the inputs_ids passed to the forward method ofXLNetModel
.d_model (
int
, optional, defaults to 1024) β Dimensionality of the encoder layers and the pooler layer.n_layer (
int
, optional, defaults to 24) β Number of hidden layers in the Transformer encoder.n_head (
int
, optional, defaults to 16) β Number of attention heads for each attention layer in the Transformer encoder.d_inner (
int
, optional, defaults to 4096) β Dimensionality of the βintermediateβ (i.e., feed-forward) layer in the Transformer encoder.ff_activation (
string
, optional, defaults to βgeluβ) β The non-linear activation function (function or string) in the encoder and pooler. If string, βgeluβ, βreluβ and βswishβ are supported.untie_r (
boolean
, optional, defaults toTrue
) β Untie relative position biasesattn_type (
string
, optional, defaults to βbiβ) β The attention type used by the model. Set βbiβ for XLNet, βuniβ for Transformer-XL.initializer_range (
float
, optional, defaults to 0.02) β The standard deviation of the truncated_normal_initializer for initializing all weight matrices.layer_norm_eps (
float
, optional, defaults to 1e-12) β The epsilon used by the layer normalization layers.dropout (
float
, optional, defaults to 0.1) β The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.mem_len (
int
orNone
, optional) β The number of tokens to cache. The key/value pairs that have already been pre-computed in a previous forward pass wonβt be re-computed. See the quickstart for more information.reuse_len (
int
orNone
, optional) β The number of tokens in the current batch to be cached and reused in the future.bi_data (
boolean
, optional, defaults toFalse
) β Whether to use bidirectional input pipeline. Usually set to True during pretraining and False during finetuning.clamp_len (
int
, optional, defaults to -1) β Clamp all relative distances larger than clamp_len. Setting this attribute to -1 means no clamping.same_length (
boolean
, optional, defaults toFalse
) β Whether to use the same attention length for each token.summary_type (
string
, optional, defaults to βlastβ) βArgument used when doing sequence summary. Used in for the multiple choice head in
XLNetForSequenceClassification
andXLNetForMultipleChoice
. Is one of the following options:βlastβ => take the last token hidden state (like XLNet)
βfirstβ => take the first token hidden state (like Bert)
βmeanβ => take the mean of all tokens hidden states
βcls_indexβ => supply a Tensor of classification token position (GPT/GPT-2)
βattnβ => Not implemented now, use multi-head attention
summary_use_proj (
boolean
, optional, defaults toTrue
) β Argument used when doing sequence summary. Used in for the multiple choice head inXLNetForSequenceClassification
andXLNetForMultipleChoice
. Add a projection after the vector extractionsummary_activation (
string
orNone
, optional) β Argument used when doing sequence summary. Used in for the multiple choice head inXLNetForSequenceClassification
andXLNetForMultipleChoice
. βtanhβ => add a tanh activation to the output, Other => no activation.summary_proj_to_labels (
boolean
, optional, defaults toTrue
) β Argument used when doing sequence summary. Used in for the multiple choice head inXLNetForSequenceClassification
andXLNetForMultipleChoice
. If True, the projection outputs to config.num_labels classes (otherwise to hidden_size). Default: False.summary_last_dropout (
float
, optional, defaults to 0.1) β Argument used when doing sequence summary. Used in for the multiple choice head inXLNetForSequenceClassification
andXLNetForMultipleChoice
. Add a dropout after the projection and activationstart_n_top (
int
, optional, defaults to 5) β Used in the SQuAD evaluation script for XLM and XLNet.end_n_top (
int
, optional, defaults to 5) β Used in the SQuAD evaluation script for XLM and XLNet.use_cache (
bool
, optional, defaults toTrue
) βWhether or not the model should return the last pre-computed hidden states.
Note
This flag behaves differently from with other models: it just controls the inference behavior, during training the model always uses
use_cache=True
.
Example:
>>> from transformers import XLNetConfig, XLNetModel >>> # Initializing a XLNet configuration >>> configuration = XLNetConfig() >>> # Initializing a model from the configuration >>> model = XLNetModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config
XLNetTokenizerΒΆ
-
class
transformers.
XLNetTokenizer
(vocab_file, do_lower_case=False, remove_space=True, keep_accents=False, bos_token='<s>', eos_token='</s>', unk_token='<unk>', sep_token='<sep>', pad_token='<pad>', cls_token='<cls>', mask_token='<mask>', additional_special_tokens=['<eop>', '<eod>'], **kwargs)[source]ΒΆ Constructs an XLNet tokenizer. Based on SentencePiece
This tokenizer inherits from
PreTrainedTokenizer
which contains most of the methods. Users should refer to the superclass for more information regarding methods.- Parameters
vocab_file (
string
) β SentencePiece file (generally has a .spm extension) that contains the vocabulary necessary to instantiate a tokenizer.do_lower_case (
bool
, optional, defaults toTrue
) β Whether to lowercase the input when tokenizing.remove_space (
bool
, optional, defaults toTrue
) β Whether to strip the text when tokenizing (removing excess spaces before and after the string).keep_accents (
bool
, optional, defaults toFalse
) β Whether to keep accents when tokenizing.bos_token (
string
, optional, defaults to β<s>β) βThe beginning of sequence token that was used during pre-training. Can be used a sequence classifier token.
Note
When building a sequence using special tokens, this is not the token that is used for the beginning of sequence. The token used is the
cls_token
.eos_token (
string
, optional, defaults to β</s>β) βThe end of sequence token.
Note
When building a sequence using special tokens, this is not the token that is used for the end of sequence. The token used is the
sep_token
.unk_token (
string
, optional, defaults to β<unk>β) β 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 (
string
, optional, defaults to β<sep>β) β The separator token, which is used when building a sequence from multiple sequences, e.g. two sequences for sequence classification or for a text and a question for question answering. It is also used as the last token of a sequence built with special tokens.pad_token (
string
, optional, defaults to β<pad>β) β The token used for padding, for example when batching sequences of different lengths.cls_token (
string
, optional, defaults to β<cls>β) β The classifier token which is used when doing sequence classification (classification of the whole sequence instead of per-token classification). It is the first token of the sequence when built with special tokens.mask_token (
string
, optional, defaults to β<mask>β) β 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.additional_special_tokens (
List[str]
, optional, defaults to["<eop>", "<eod>"]
) β Additional special tokens used by the tokenizer.
-
sp_model
ΒΆ The SentencePiece processor that is used for every conversion (string, tokens and IDs).
- Type
SentencePieceProcessor
-
build_inputs_with_special_tokens
(token_ids_0: List[int], token_ids_1: Optional[List[int]] = None) → List[int][source]ΒΆ Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens. An XLNet sequence has the following format:
single sequence:
X <sep> <cls>
pair of sequences:
A <sep> B <sep> <cls>
- Parameters
token_ids_0 (
List[int]
) β List of IDs to which the special tokens will be addedtoken_ids_1 (
List[int]
, optional) β Optional second list of IDs for sequence pairs.
- Returns
list of input IDs with the appropriate special tokens.
- Return type
List[int]
-
create_token_type_ids_from_sequences
(token_ids_0: List[int], token_ids_1: Optional[List[int]] = None) → List[int][source]ΒΆ Creates a mask from the two sequences passed to be used in a sequence-pair classification task. An XLNet sequence pair mask has the following format: 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 2 | first sequence | second sequence | CLS segment ID
if token_ids_1 is None, only returns the first portion of the mask (0βs).
- Parameters
token_ids_0 (
List[int]
) β List of ids.token_ids_1 (
List[int]
, optional) β Optional second list of IDs for sequence pairs.
- Returns
List of token type IDs according to the given sequence(s).
- Return type
List[int]
-
get_special_tokens_mask
(token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False) → List[int][source]ΒΆ Retrieves sequence ids from a token list that has no special tokens added. This method is called when adding special tokens using the tokenizer
prepare_for_model
methods.- Parameters
token_ids_0 (
List[int]
) β List of ids.token_ids_1 (
List[int]
, optional) β Optional second list of IDs for sequence pairs.already_has_special_tokens (
bool
, optional, defaults toFalse
) β Set to True if the token list is already formatted with special tokens for the model
- Returns
A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
- Return type
List[int]
XLNet specific outputsΒΆ
-
class
transformers.modeling_xlnet.
XLNetModelOutput
(last_hidden_state: torch.FloatTensor, mems: Optional[List[torch.FloatTensor]] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]ΒΆ Output type of
XLNetModel
.- Parameters
last_hidden_state (
torch.FloatTensor
of shape(batch_size, num_predict, hidden_size)
) βSequence of hidden-states at the last layer of the model.
num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
isNone
, thennum_predict
corresponds tosequence_length
.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
torch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple 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.
-
class
transformers.modeling_xlnet.
XLNetLMHeadModelOutput
(loss: Optional[torch.FloatTensor] = None, logits: torch.FloatTensor = None, mems: Optional[List[torch.FloatTensor]] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]ΒΆ Output type of
XLNetLMHeadModel
.- Parameters
loss (
torch.FloatTensor
of shape (1,), optional, returned whenlabels
is provided) β Language modeling loss (for next-token prediction).logits (
torch.FloatTensor
of shape(batch_size, num_predict, config.vocab_size)
) βPrediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
isNone
, thennum_predict
corresponds tosequence_length
.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
torch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple 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.
-
class
transformers.modeling_xlnet.
XLNetForSequenceClassificationOutput
(loss: Optional[torch.FloatTensor] = None, logits: torch.FloatTensor = None, mems: Optional[List[torch.FloatTensor]] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]ΒΆ Output type of
XLNetForSequenceClassification
.- Parameters
loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabel
is provided) β Classification (or regression if config.num_labels==1) loss.logits (
torch.FloatTensor
of shape(batch_size, config.num_labels)
) β Classification (or regression if config.num_labels==1) scores (before SoftMax).mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
torch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple 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.
-
class
transformers.modeling_xlnet.
XLNetForMultipleChoiceOutput
(loss: Optional[torch.FloatTensor] = None, logits: torch.FloatTensor = None, mems: Optional[List[torch.FloatTensor]] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]ΒΆ Output type of
XLNetForMultipleChoice
.- Parameters
loss (
torch.FloatTensor
of shape (1,), optional, returned whenlabels
is provided) β Classification loss.logits (
torch.FloatTensor
of shape(batch_size, num_choices)
) βnum_choices is the second dimension of the input tensors. (see input_ids above).
Classification scores (before SoftMax).
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
torch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple 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.
-
class
transformers.modeling_xlnet.
XLNetForTokenClassificationOutput
(loss: Optional[torch.FloatTensor] = None, logits: torch.FloatTensor = None, mems: Optional[List[torch.FloatTensor]] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]ΒΆ Output type of
XLNetForTokenClassificationOutput
.- Parameters
loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabels
is provided) β Classification loss.logits (
torch.FloatTensor
of shape(batch_size, sequence_length, config.num_labels)
) β Classification scores (before SoftMax).mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
torch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple 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.
-
class
transformers.modeling_xlnet.
XLNetForQuestionAnsweringSimpleOutput
(loss: Optional[torch.FloatTensor] = None, start_logits: torch.FloatTensor = None, end_logits: torch.FloatTensor = None, mems: Optional[List[torch.FloatTensor]] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]ΒΆ Output type of
XLNetForQuestionAnsweringSimple
.- Parameters
loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabels
is provided) β Total span extraction loss is the sum of a Cross-Entropy for the start and end positions.start_logits (
torch.FloatTensor
of shape(batch_size, sequence_length,)
) β Span-start scores (before SoftMax).end_logits (
torch.FloatTensor
of shape(batch_size, sequence_length,)
) β Span-end scores (before SoftMax).mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
torch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple 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.
-
class
transformers.modeling_xlnet.
XLNetForQuestionAnsweringOutput
(loss: Optional[torch.FloatTensor] = None, start_top_log_probs: Optional[torch.FloatTensor] = None, start_top_index: Optional[torch.LongTensor] = None, end_top_log_probs: Optional[torch.FloatTensor] = None, end_top_index: Optional[torch.LongTensor] = None, cls_logits: Optional[torch.FloatTensor] = None, mems: Optional[List[torch.FloatTensor]] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]ΒΆ Output type of
XLNetForQuestionAnswering
.- Parameters
loss (
torch.FloatTensor
of shape(1,)
, optional, returned if bothstart_positions
andend_positions
are provided) β Classification loss as the sum of start token, end token (and is_impossible if provided) classification losses.start_top_log_probs (
torch.FloatTensor
of shape(batch_size, config.start_n_top)
, optional, returned ifstart_positions
orend_positions
is not provided) β Log probabilities for the top config.start_n_top start token possibilities (beam-search).start_top_index (
torch.LongTensor
of shape(batch_size, config.start_n_top)
, optional, returned ifstart_positions
orend_positions
is not provided) β Indices for the top config.start_n_top start token possibilities (beam-search).end_top_log_probs (
torch.FloatTensor
of shape(batch_size, config.start_n_top * config.end_n_top)
, optional, returned ifstart_positions
orend_positions
is not provided) β Log probabilities for the topconfig.start_n_top * config.end_n_top
end token possibilities (beam-search).end_top_index (
torch.LongTensor
of shape(batch_size, config.start_n_top * config.end_n_top)
, optional, returned ifstart_positions
orend_positions
is not provided) β Indices for the topconfig.start_n_top * config.end_n_top
end token possibilities (beam-search).cls_logits (
torch.FloatTensor
of shape(batch_size,)
, optional, returned ifstart_positions
orend_positions
is not provided) β Log probabilities for theis_impossible
label of the answers.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
torch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple 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.
-
class
transformers.modeling_tf_xlnet.
TFXLNetModelOutput
(last_hidden_state: tensorflow.python.framework.ops.Tensor = None, mems: Optional[List[tensorflow.python.framework.ops.Tensor]] = None, hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)[source]ΒΆ Output type of
TFXLNetModel
.- Parameters
last_hidden_state (
tf.Tensor
of shape(batch_size, num_predict, hidden_size)
) βSequence of hidden-states at the last layer of the model.
num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
isNone
, thennum_predict
corresponds tosequence_length
.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
tf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple of
tf.Tensor
(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.
-
class
transformers.modeling_tf_xlnet.
TFXLNetLMHeadModelOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: tensorflow.python.framework.ops.Tensor = None, mems: Optional[List[tensorflow.python.framework.ops.Tensor]] = None, hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)[source]ΒΆ Output type of
TFXLNetLMHeadModel
.- Parameters
loss (
tf.Tensor
of shape (1,), optional, returned whenlabels
is provided) β Language modeling loss (for next-token prediction).logits (
tf.Tensor
of shape(batch_size, num_predict, config.vocab_size)
) βPrediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
isNone
, thennum_predict
corresponds tosequence_length
.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
tf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple of
tf.Tensor
(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.
-
class
transformers.modeling_tf_xlnet.
TFXLNetForSequenceClassificationOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: tensorflow.python.framework.ops.Tensor = None, mems: Optional[List[tensorflow.python.framework.ops.Tensor]] = None, hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)[source]ΒΆ Output type of
TFXLNetForSequenceClassification
.- Parameters
loss (
tf.Tensor
of shape(1,)
, optional, returned whenlabel
is provided) β Classification (or regression if config.num_labels==1) loss.logits (
tf.Tensor
of shape(batch_size, config.num_labels)
) β Classification (or regression if config.num_labels==1) scores (before SoftMax).mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
tf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple of
tf.Tensor
(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.
-
class
transformers.modeling_tf_xlnet.
TFXLNetForMultipleChoiceOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: tensorflow.python.framework.ops.Tensor = None, mems: Optional[List[tensorflow.python.framework.ops.Tensor]] = None, hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)[source]ΒΆ Output type of
TFXLNetForMultipleChoice
.- Parameters
loss (
tf.Tensor
of shape (1,), optional, returned whenlabels
is provided) β Classification loss.logits (
tf.Tensor
of shape(batch_size, num_choices)
) βnum_choices is the second dimension of the input tensors. (see input_ids above).
Classification scores (before SoftMax).
mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
tf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple of
tf.Tensor
(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.
-
class
transformers.modeling_tf_xlnet.
TFXLNetForTokenClassificationOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: tensorflow.python.framework.ops.Tensor = None, mems: Optional[List[tensorflow.python.framework.ops.Tensor]] = None, hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)[source]ΒΆ Output type of
TFXLNetForTokenClassificationOutput
.- Parameters
loss (
tf.Tensor
of shape(1,)
, optional, returned whenlabels
is provided) β Classification loss.logits (
tf.Tensor
of shape(batch_size, sequence_length, config.num_labels)
) β Classification scores (before SoftMax).mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
tf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple of
tf.Tensor
(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.
-
class
transformers.modeling_tf_xlnet.
TFXLNetForQuestionAnsweringSimpleOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, start_logits: tensorflow.python.framework.ops.Tensor = None, end_logits: tensorflow.python.framework.ops.Tensor = None, mems: Optional[List[tensorflow.python.framework.ops.Tensor]] = None, hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)[source]ΒΆ Output type of
TFXLNetForQuestionAnsweringSimple
.- Parameters
loss (
tf.Tensor
of shape(1,)
, optional, returned whenlabels
is provided) β Total span extraction loss is the sum of a Cross-Entropy for the start and end positions.start_logits (
tf.Tensor
of shape(batch_size, sequence_length,)
) β Span-start scores (before SoftMax).end_logits (
tf.Tensor
of shape(batch_size, sequence_length,)
) β Span-end scores (before SoftMax).mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) βTuple of
tf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) βTuple of
tf.Tensor
(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.
XLNetModelΒΆ
-
class
transformers.
XLNetModel
(config)[source]ΒΆ The bare XLNet Model transformer outputting raw hidden-states without any specific head on top.
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
XLNetConfig
) β 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.
-
create_mask
(qlen, mlen)[source]ΒΆ Creates causal attention mask. Float mask where 1.0 indicates masked, 0.0 indicates not-masked.
- Parameters
qlen β Sequence length
mlen β Mask length
same_length=False: same_length=True: <mlen > < qlen > <mlen > < qlen > ^ [0 0 0 0 0 1 1 1 1] [0 0 0 0 0 1 1 1 1] [0 0 0 0 0 0 1 1 1] [1 0 0 0 0 0 1 1 1] qlen [0 0 0 0 0 0 0 1 1] [1 1 0 0 0 0 0 1 1] [0 0 0 0 0 0 0 0 1] [1 1 1 0 0 0 0 0 1] v [0 0 0 0 0 0 0 0 0] [1 1 1 1 0 0 0 0 0]
-
forward
(input_ids=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]ΒΆ The
XLNetModel
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.BertTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed. use_cache has to be set to True to make use of mems.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
torch.FloatTensor
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B token. The classifier token should be represented by a2
.input_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.
- Returns
A
XLNetModelOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftorch.FloatTensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.last_hidden_state (
torch.FloatTensor
of shape(batch_size, num_predict, hidden_size)
) β Sequence of hidden-states at the last layer of the model.num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
isNone
, thennum_predict
corresponds tosequence_length
.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β 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-attention heads.
- Return type
XLNetModelOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import XLNetTokenizer, XLNetModel >>> import torch >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = XLNetModel.from_pretrained('xlnet-base-cased', return_dict=True) >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> outputs = model(**inputs) >>> last_hidden_states = outputs.last_hidden_state
XLNetLMHeadModelΒΆ
-
class
transformers.
XLNetLMHeadModel
(config)[source]ΒΆ XLNet Model with a language modeling head on top (linear layer with weights tied to the input embeddings).
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
XLNetConfig
) β 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.
-
forward
(input_ids=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, labels=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]ΒΆ The
XLNetLMHeadModel
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.BertTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed. use_cache has to be set to True to make use of mems.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
torch.FloatTensor
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B token. The classifier token should be represented by a2
.input_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) βIf set to
True
, the model will return aModelOutput
instead of a plain tuple.- labels (
torch.LongTensor
of shape(batch_size, num_predict)
, optional): Labels for masked language modeling. num_predict corresponds to target_mapping.shape[1]. If target_mapping is None, then num_predict corresponds to sequence_length. The labels should correspond to the masked input words that should be predicted and depends on target_mapping. Note in order to perform standard auto-regressive language modeling a <mask> token has to be added to the input_ids (see prepare_inputs_for_generation fn and examples below) Indices are selected in
[-100, 0, ..., config.vocab_size]
All labels set to-100
are ignored, the loss is only computed for labels in[0, ..., config.vocab_size]
- labels (
- Returns
A
XLNetLMHeadModelOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftorch.FloatTensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
torch.FloatTensor
of shape (1,), optional, returned whenlabels
is provided) Language modeling loss (for next-token prediction).logits (
torch.FloatTensor
of shape(batch_size, num_predict, config.vocab_size)
) β Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
isNone
, thennum_predict
corresponds tosequence_length
.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β 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-attention heads.
Examples:
>>> from transformers import XLNetTokenizer, XLNetLMHeadModel >>> import torch >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-large-cased') >>> model = XLNetLMHeadModel.from_pretrained('xlnet-large-cased', return_dict=True) >>> # We show how to setup inputs to predict a next token using a bi-directional context. >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is very <mask>", add_special_tokens=False)).unsqueeze(0) # We will predict the masked token >>> perm_mask = torch.zeros((1, input_ids.shape[1], input_ids.shape[1]), dtype=torch.float) >>> perm_mask[:, :, -1] = 1.0 # Previous tokens don't see last token >>> target_mapping = torch.zeros((1, 1, input_ids.shape[1]), dtype=torch.float) # Shape [1, 1, seq_length] => let's predict one token >>> target_mapping[0, 0, -1] = 1.0 # Our first (and only) prediction will be the last token of the sequence (the masked token) >>> outputs = model(input_ids, perm_mask=perm_mask, target_mapping=target_mapping) >>> next_token_logits = outputs[0] # Output has shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size] >>> # The same way can the XLNetLMHeadModel be used to be trained by standard auto-regressive language modeling. >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is very <mask>", add_special_tokens=False)).unsqueeze(0) # We will predict the masked token >>> labels = torch.tensor(tokenizer.encode("cute", add_special_tokens=False)).unsqueeze(0) >>> assert labels.shape[0] == 1, 'only one word will be predicted' >>> perm_mask = torch.zeros((1, input_ids.shape[1], input_ids.shape[1]), dtype=torch.float) >>> perm_mask[:, :, -1] = 1.0 # Previous tokens don't see last token as is done in standard auto-regressive lm training >>> target_mapping = torch.zeros((1, 1, input_ids.shape[1]), dtype=torch.float) # Shape [1, 1, seq_length] => let's predict one token >>> target_mapping[0, 0, -1] = 1.0 # Our first (and only) prediction will be the last token of the sequence (the masked token) >>> outputs = model(input_ids, perm_mask=perm_mask, target_mapping=target_mapping, labels=labels) >>> loss = outputs.loss >>> next_token_logits = outputs.logits # Logits have shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
- Return type
XLNetLMHeadModelOutput
ortuple(torch.FloatTensor)
XLNetForSequenceClassificationΒΆ
-
class
transformers.
XLNetForSequenceClassification
(config)[source]ΒΆ XLNet Model with a sequence classification/regression head on top (a linear layer on top of the pooled output) e.g. for GLUE tasks.
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
XLNetConfig
) β 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.
-
forward
(input_ids=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, labels=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]ΒΆ The
XLNetForSequenceClassification
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.BertTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed. use_cache has to be set to True to make use of mems.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
torch.FloatTensor
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B token. The classifier token should be represented by a2
.input_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.labels (
torch.LongTensor
of shape(batch_size,)
, optional) β Labels for computing the sequence classification/regression loss. Indices should be in[0, ..., config.num_labels - 1]
. Ifconfig.num_labels == 1
a regression loss is computed (Mean-Square loss), Ifconfig.num_labels > 1
a classification loss is computed (Cross-Entropy).
- Returns
A
XLNetForSequenceClassificationOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftorch.FloatTensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabel
is provided) β Classification (or regression if config.num_labels==1) loss.logits (
torch.FloatTensor
of shape(batch_size, config.num_labels)
) β Classification (or regression if config.num_labels==1) scores (before SoftMax).mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β 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-attention heads.
- Return type
XLNetForSequenceClassificationOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import XLNetTokenizer, XLNetForSequenceClassification >>> import torch >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = XLNetForSequenceClassification.from_pretrained('xlnet-base-cased', return_dict=True) >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> labels = torch.tensor([1]).unsqueeze(0) # Batch size 1 >>> outputs = model(**inputs, labels=labels) >>> loss = outputs.loss >>> logits = outputs.logits
XLNetForMultipleChoiceΒΆ
-
class
transformers.
XLNetForMultipleChoice
(config)[source]ΒΆ XLNet Model with a multiple choice classification head on top (a linear layer on top of the pooled output and a softmax) e.g. for RACE/SWAG tasks.
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
XLNetConfig
) β 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.
-
forward
(input_ids=None, token_type_ids=None, input_mask=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, head_mask=None, inputs_embeds=None, labels=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]ΒΆ The
XLNetForMultipleChoice
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, num_choices, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.BertTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shape(batch_size, num_choices, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed. use_cache has to be set to True to make use of mems.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
torch.FloatTensor
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
torch.LongTensor
of shape(batch_size, num_choices, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B token. The classifier token should be represented by a2
.input_mask (
torch.FloatTensor
of shape(batch_size, num_choices, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.labels (
torch.LongTensor
of shape(batch_size,)
, optional) β Labels for computing the multiple choice classification loss. Indices should be in[0, ..., num_choices]
where num_choices is the size of the second dimension of the input tensors. (see input_ids above)
- Returns
A
XLNetForMultipleChoiceOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftorch.FloatTensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
torch.FloatTensor
of shape (1,), optional, returned whenlabels
is provided) β Classification loss.logits (
torch.FloatTensor
of shape(batch_size, num_choices)
) β num_choices is the second dimension of the input tensors. (see input_ids above).Classification scores (before SoftMax).
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β 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-attention heads.
- Return type
XLNetForMultipleChoiceOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import XLNetTokenizer, XLNetForMultipleChoice >>> import torch >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = XLNetForMultipleChoice.from_pretrained('xlnet-base-cased', return_dict=True) >>> prompt = "In Italy, pizza served in formal settings, such as at a restaurant, is presented unsliced." >>> choice0 = "It is eaten with a fork and a knife." >>> choice1 = "It is eaten while held in the hand." >>> labels = torch.tensor(0).unsqueeze(0) # choice0 is correct (according to Wikipedia ;)), batch size 1 >>> encoding = tokenizer([[prompt, prompt], [choice0, choice1]], return_tensors='pt', padding=True) >>> outputs = model(**{k: v.unsqueeze(0) for k,v in encoding.items()}, labels=labels) # batch size is 1 >>> # the linear classifier still needs to be trained >>> loss = outputs.loss >>> logits = outputs.logits
XLNetForTokenClassificationΒΆ
-
class
transformers.
XLNetForTokenClassification
(config)[source]ΒΆ XLNet Model with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for Named-Entity-Recognition (NER) tasks.
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
XLNetConfig
) β 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.
-
forward
(input_ids=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, labels=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]ΒΆ The
XLNetForTokenClassification
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.BertTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed. use_cache has to be set to True to make use of mems.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
torch.FloatTensor
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B token. The classifier token should be represented by a2
.input_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.labels (
torch.LongTensor
of shape(batch_size,)
, optional) β Labels for computing the multiple choice classification loss. Indices should be in[0, ..., num_choices]
where num_choices is the size of the second dimension of the input tensors. (see input_ids above)
- Returns
A
XLNetForTokenClassificationOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftorch.FloatTensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabels
is provided) β Classification loss.logits (
torch.FloatTensor
of shape(batch_size, sequence_length, config.num_labels)
) β Classification scores (before SoftMax).mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β 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-attention heads.
- Return type
XLNetForTokenClassificationOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import XLNetTokenizer, XLNetForTokenClassification >>> import torch >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = XLNetForTokenClassification.from_pretrained('xlnet-base-cased', return_dict=True) >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> labels = torch.tensor([1] * inputs["input_ids"].size(1)).unsqueeze(0) # Batch size 1 >>> outputs = model(**inputs, labels=labels) >>> loss = outputs.loss >>> logits = outputs.logits
XLNetForQuestionAnsweringSimpleΒΆ
-
class
transformers.
XLNetForQuestionAnsweringSimple
(config)[source]ΒΆ XLNet Model with a span classification head on top for extractive question-answering tasks like SQuAD (a linear layers on top of the hidden-states output to compute span start logits and span end logits).
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
XLNetConfig
) β 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.
-
forward
(input_ids=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, start_positions=None, end_positions=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]ΒΆ The
XLNetForQuestionAnsweringSimple
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.BertTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed. use_cache has to be set to True to make use of mems.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
torch.FloatTensor
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B token. The classifier token should be represented by a2
.input_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.start_positions (
torch.LongTensor
of shape(batch_size,)
, optional) β Labels for position (index) of the start of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length). Position outside of the sequence are not taken into account for computing the loss.end_positions (
torch.LongTensor
of shape(batch_size,)
, optional) β Labels for position (index) of the end of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length). Position outside of the sequence are not taken into account for computing the loss.
- Returns
A
XLNetForQuestionAnsweringSimpleOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftorch.FloatTensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabels
is provided) β Total span extraction loss is the sum of a Cross-Entropy for the start and end positions.start_logits (
torch.FloatTensor
of shape(batch_size, sequence_length,)
) β Span-start scores (before SoftMax).end_logits (
torch.FloatTensor
of shape(batch_size, sequence_length,)
) β Span-end scores (before SoftMax).mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β 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-attention heads.
- Return type
XLNetForQuestionAnsweringSimpleOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import XLNetTokenizer, XLNetForQuestionAnsweringSimple >>> import torch >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = XLNetForQuestionAnsweringSimple.from_pretrained('xlnet-base-cased', return_dict=True) >>> question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet" >>> inputs = tokenizer(question, text, return_tensors='pt') >>> start_positions = torch.tensor([1]) >>> end_positions = torch.tensor([3]) >>> outputs = model(**inputs, start_positions=start_positions, end_positions=end_positions) >>> loss = outputs.loss >>> start_scores = outputs.start_logits >>> end_scores = outputs.end_logits
XLNetForQuestionAnsweringΒΆ
-
class
transformers.
XLNetForQuestionAnswering
(config)[source]ΒΆ XLNet Model with a span classification head on top for extractive question-answering tasks like SQuAD (a linear layers on top of the hidden-states output to compute span start logits and span end logits).
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
XLNetConfig
) β 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.
-
forward
(input_ids=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, start_positions=None, end_positions=None, is_impossible=None, cls_index=None, p_mask=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]ΒΆ The
XLNetForQuestionAnswering
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.BertTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed. use_cache has to be set to True to make use of mems.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
torch.FloatTensor
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B token. The classifier token should be represented by a2
.input_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) βIf set to
True
, the model will return aModelOutput
instead of a plain tuple.- start_positions (
torch.LongTensor
of shape(batch_size,)
, optional): Labels for position (index) of the start of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length). Position outside of the sequence are not taken into account for computing the loss.
- end_positions (
torch.LongTensor
of shape(batch_size,)
, optional): Labels for position (index) of the end of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length). Position outside of the sequence are not taken into account for computing the loss.
- is_impossible (
torch.LongTensor
of shape(batch_size,)
, optional): Labels whether a question has an answer or no answer (SQuAD 2.0)
- cls_index (
torch.LongTensor
of shape(batch_size,)
, optional): Labels for position (index) of the classification token to use as input for computing plausibility of the answer.
- p_mask (
torch.FloatTensor
of shape(batch_size, sequence_length)
, optional): Optional mask of tokens which canβt be in answers (e.g. [CLS], [PAD], β¦). 1.0 means token should be masked. 0.0 mean token is not masked.
- start_positions (
- Returns
A
XLNetForQuestionAnsweringOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftorch.FloatTensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
torch.FloatTensor
of shape(1,)
, optional, returned if bothstart_positions
andend_positions
are provided) β Classification loss as the sum of start token, end token (and is_impossible if provided) classification losses.start_top_log_probs (
torch.FloatTensor
of shape(batch_size, config.start_n_top)
, optional, returned ifstart_positions
orend_positions
is not provided) β Log probabilities for the top config.start_n_top start token possibilities (beam-search).start_top_index (
torch.LongTensor
of shape(batch_size, config.start_n_top)
, optional, returned ifstart_positions
orend_positions
is not provided) β Indices for the top config.start_n_top start token possibilities (beam-search).end_top_log_probs (
torch.FloatTensor
of shape(batch_size, config.start_n_top * config.end_n_top)
, optional, returned ifstart_positions
orend_positions
is not provided) β Log probabilities for the topconfig.start_n_top * config.end_n_top
end token possibilities (beam-search).end_top_index (
torch.LongTensor
of shape(batch_size, config.start_n_top * config.end_n_top)
, optional, returned ifstart_positions
orend_positions
is not provided) β Indices for the topconfig.start_n_top * config.end_n_top
end token possibilities (beam-search).cls_logits (
torch.FloatTensor
of shape(batch_size,)
, optional, returned ifstart_positions
orend_positions
is not provided) β Log probabilities for theis_impossible
label of the answers.mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β 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-attention heads.
Example:
>>> from transformers import XLNetTokenizer, XLNetForQuestionAnswering >>> import torch >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = XLNetForQuestionAnswering.from_pretrained('xlnet-base-cased', return_dict=True) >>> input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True)).unsqueeze(0) # Batch size 1 >>> start_positions = torch.tensor([1]) >>> end_positions = torch.tensor([3]) >>> outputs = model(input_ids, start_positions=start_positions, end_positions=end_positions) >>> loss = outputs.loss
- Return type
XLNetForQuestionAnsweringOutput
ortuple(torch.FloatTensor)
TFXLNetModelΒΆ
-
class
transformers.
TFXLNetModel
(*args, **kwargs)[source]ΒΆ The bare XLNet Model transformer outputing raw hidden-states without any specific head on top.
Note
TF 2.0 models accepts two formats as inputs:
having all inputs as keyword arguments (like PyTorch models), or
having all inputs as a list, tuple or dict in the first positional arguments.
This second option is useful when using
tf.keras.Model.fit()
method which currently requires having all the tensors in the first argument of the model call function:model(inputs)
.If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :
a single Tensor with input_ids only and nothing else:
model(inputs_ids)
a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
model([input_ids, attention_mask])
ormodel([input_ids, attention_mask, token_type_ids])
a dictionary with one or several input Tensors associated to the input names given in the docstring:
model({'input_ids': input_ids, 'token_type_ids': token_type_ids})
- Parameters
config (
XLNetConfig
) β 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.
-
call
(inputs, **kwargs)[source]ΒΆ The
TFXLNetModel
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.XLNetTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
tf.Tensor
orNumpy array
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B tokeninput_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
tf.Tensor
orNumpy array
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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 (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.
- Returns
A
TFXLNetModelOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftf.Tensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.last_hidden_state (
tf.Tensor
of shape(batch_size, num_predict, hidden_size)
) β Sequence of hidden-states at the last layer of the model.num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
isNone
, thennum_predict
corresponds tosequence_length
.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β Tuple oftf.Tensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
TFXLNetModelOutput
ortuple(tf.Tensor)
Example:
>>> from transformers import XLNetTokenizer, TFXLNetModel >>> import tensorflow as tf >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = TFXLNetModel.from_pretrained('xlnet-base-cased') >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="tf") >>> outputs = model(inputs) >>> last_hidden_states = outputs[0] # The last hidden-state is the first element of the output tuple
TFXLNetLMHeadModelΒΆ
-
class
transformers.
TFXLNetLMHeadModel
(*args, **kwargs)[source]ΒΆ XLNet Model with a language modeling head on top (linear layer with weights tied to the input embeddings).
Note
TF 2.0 models accepts two formats as inputs:
having all inputs as keyword arguments (like PyTorch models), or
having all inputs as a list, tuple or dict in the first positional arguments.
This second option is useful when using
tf.keras.Model.fit()
method which currently requires having all the tensors in the first argument of the model call function:model(inputs)
.If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :
a single Tensor with input_ids only and nothing else:
model(inputs_ids)
a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
model([input_ids, attention_mask])
ormodel([input_ids, attention_mask, token_type_ids])
a dictionary with one or several input Tensors associated to the input names given in the docstring:
model({'input_ids': input_ids, 'token_type_ids': token_type_ids})
- Parameters
config (
XLNetConfig
) β 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.
-
call
(inputs, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, use_cache=True, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False)[source]ΒΆ The
TFXLNetLMHeadModel
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.XLNetTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
tf.Tensor
orNumpy array
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B tokeninput_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
tf.Tensor
orNumpy array
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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 (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) βIf set to
True
, the model will return aModelOutput
instead of a plain tuple.- labels (
tf.Tensor
of shape(batch_size, sequence_length)
, optional): Labels for computing the cross entropy classification loss. Indices should be in
[0, ..., config.vocab_size - 1]
.
- labels (
- Returns
A
TFXLNetLMHeadModelOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftf.Tensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
tf.Tensor
of shape (1,), optional, returned whenlabels
is provided) Language modeling loss (for next-token prediction).logits (
tf.Tensor
of shape(batch_size, num_predict, config.vocab_size)
) β Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
isNone
, thennum_predict
corresponds tosequence_length
.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β Tuple oftf.Tensor
(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.
Examples:
>>> import tensorflow as tf >>> import numpy as np >>> from transformers import XLNetTokenizer, TFXLNetLMHeadModel >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-large-cased') >>> model = TFXLNetLMHeadModel.from_pretrained('xlnet-large-cased') >>> # We show how to setup inputs to predict a next token using a bi-directional context. >>> input_ids = tf.constant(tokenizer.encode("Hello, my dog is very <mask>", add_special_tokens=True))[None, :] # We will predict the masked token >>> perm_mask = np.zeros((1, input_ids.shape[1], input_ids.shape[1])) >>> perm_mask[:, :, -1] = 1.0 # Previous tokens don't see last token >>> target_mapping = np.zeros((1, 1, input_ids.shape[1])) # Shape [1, 1, seq_length] => let's predict one token >>> target_mapping[0, 0, -1] = 1.0 # Our first (and only) prediction will be the last token of the sequence (the masked token) >>> outputs = model(input_ids, perm_mask=tf.constant(perm_mask, dtype=tf.float32), target_mapping=tf.constant(target_mapping, dtype=tf.float32)) >>> next_token_logits = outputs[0] # Output has shape [target_mapping.size(0), target_mapping.size(1), config.vocab_size]
- Return type
TFXLNetLMHeadModelOutput
ortuple(tf.Tensor)
TFXLNetForSequenceClassificationΒΆ
-
class
transformers.
TFXLNetForSequenceClassification
(*args, **kwargs)[source]ΒΆ XLNet Model with a sequence classification/regression head on top (a linear layer on top of the pooled output) e.g. for GLUE tasks.
Note
TF 2.0 models accepts two formats as inputs:
having all inputs as keyword arguments (like PyTorch models), or
having all inputs as a list, tuple or dict in the first positional arguments.
This second option is useful when using
tf.keras.Model.fit()
method which currently requires having all the tensors in the first argument of the model call function:model(inputs)
.If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :
a single Tensor with input_ids only and nothing else:
model(inputs_ids)
a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
model([input_ids, attention_mask])
ormodel([input_ids, attention_mask, token_type_ids])
a dictionary with one or several input Tensors associated to the input names given in the docstring:
model({'input_ids': input_ids, 'token_type_ids': token_type_ids})
- Parameters
config (
XLNetConfig
) β 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.
-
call
(inputs=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, use_cache=True, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False)[source]ΒΆ The
TFXLNetForSequenceClassification
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.XLNetTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
tf.Tensor
orNumpy array
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B tokeninput_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
tf.Tensor
orNumpy array
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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 (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.labels (
tf.Tensor
of shape(batch_size,)
, optional) β Labels for computing the sequence classification/regression loss. Indices should be in[0, ..., config.num_labels - 1]
. Ifconfig.num_labels == 1
a regression loss is computed (Mean-Square loss), Ifconfig.num_labels > 1
a classification loss is computed (Cross-Entropy).
- Returns
A
TFXLNetForSequenceClassificationOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftf.Tensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
tf.Tensor
of shape(1,)
, optional, returned whenlabel
is provided) β Classification (or regression if config.num_labels==1) loss.logits (
tf.Tensor
of shape(batch_size, config.num_labels)
) β Classification (or regression if config.num_labels==1) scores (before SoftMax).mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β Tuple oftf.Tensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
TFXLNetForSequenceClassificationOutput
ortuple(tf.Tensor)
Example:
>>> from transformers import XLNetTokenizer, TFXLNetForSequenceClassification >>> import tensorflow as tf >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = TFXLNetForSequenceClassification.from_pretrained('xlnet-base-cased') >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="tf") >>> inputs["labels"] = tf.reshape(tf.constant(1), (-1, 1)) # Batch size 1 >>> outputs = model(inputs) >>> loss, logits = outputs[:2]
TFLNetForMultipleChoiceΒΆ
-
class
transformers.
TFXLNetForMultipleChoice
(*args, **kwargs)[source]ΒΆ XLNET Model with a multiple choice classification head on top (a linear layer on top of the pooled output and a softmax) e.g. for RocStories/SWAG tasks.
Note
TF 2.0 models accepts two formats as inputs:
having all inputs as keyword arguments (like PyTorch models), or
having all inputs as a list, tuple or dict in the first positional arguments.
This second option is useful when using
tf.keras.Model.fit()
method which currently requires having all the tensors in the first argument of the model call function:model(inputs)
.If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :
a single Tensor with input_ids only and nothing else:
model(inputs_ids)
a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
model([input_ids, attention_mask])
ormodel([input_ids, attention_mask, token_type_ids])
a dictionary with one or several input Tensors associated to the input names given in the docstring:
model({'input_ids': input_ids, 'token_type_ids': token_type_ids})
- Parameters
config (
XLNetConfig
) β 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.
-
call
(inputs=None, token_type_ids=None, input_mask=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, head_mask=None, inputs_embeds=None, use_cache=True, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False)[source]ΒΆ The
TFXLNetForMultipleChoice
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.XLNetTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
tf.Tensor
orNumpy array
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B tokeninput_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
tf.Tensor
orNumpy array
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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 (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.labels (
tf.Tensor
of shape(batch_size,)
, optional) β Labels for computing the multiple choice classification loss. Indices should be in[0, ..., num_choices]
where num_choices is the size of the second dimension of the input tensors. (see input_ids above)
- Returns
A
TFXLNetForMultipleChoiceOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftf.Tensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
tf.Tensor
of shape (1,), optional, returned whenlabels
is provided) β Classification loss.logits (
tf.Tensor
of shape(batch_size, num_choices)
) β num_choices is the second dimension of the input tensors. (see input_ids above).Classification scores (before SoftMax).
mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β Tuple oftf.Tensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
TFXLNetForMultipleChoiceOutput
ortuple(tf.Tensor)
Example:
>>> from transformers import XLNetTokenizer, TFXLNetForMultipleChoice >>> import tensorflow as tf >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = TFXLNetForMultipleChoice.from_pretrained('xlnet-base-cased') >>> prompt = "In Italy, pizza served in formal settings, such as at a restaurant, is presented unsliced." >>> choice0 = "It is eaten with a fork and a knife." >>> choice1 = "It is eaten while held in the hand." >>> encoding = tokenizer([[prompt, prompt], [choice0, choice1]], return_tensors='tf', padding=True) >>> inputs = {k: tf.expand_dims(v, 0) for k, v in encoding.items()} >>> outputs = model(inputs) # batch size is 1 >>> # the linear classifier still needs to be trained >>> logits = outputs[0]
-
property
dummy_inputs
ΒΆ Dummy inputs to build the network.
- Returns
tf.Tensor with dummy inputs
TFXLNetForTokenClassificationΒΆ
-
class
transformers.
TFXLNetForTokenClassification
(*args, **kwargs)[source]ΒΆ XLNet Model with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for Named-Entity-Recognition (NER) tasks.
Note
TF 2.0 models accepts two formats as inputs:
having all inputs as keyword arguments (like PyTorch models), or
having all inputs as a list, tuple or dict in the first positional arguments.
This second option is useful when using
tf.keras.Model.fit()
method which currently requires having all the tensors in the first argument of the model call function:model(inputs)
.If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :
a single Tensor with input_ids only and nothing else:
model(inputs_ids)
a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
model([input_ids, attention_mask])
ormodel([input_ids, attention_mask, token_type_ids])
a dictionary with one or several input Tensors associated to the input names given in the docstring:
model({'input_ids': input_ids, 'token_type_ids': token_type_ids})
- Parameters
config (
XLNetConfig
) β 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.
-
call
(inputs=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, use_cache=True, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False)[source]ΒΆ The
TFXLNetForTokenClassification
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.XLNetTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
tf.Tensor
orNumpy array
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B tokeninput_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
tf.Tensor
orNumpy array
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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 (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.labels (
tf.Tensor
of shape(batch_size, sequence_length)
, optional) β Labels for computing the token classification loss. Indices should be in[0, ..., config.num_labels - 1]
.
- Returns
A
TFXLNetForTokenClassificationOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftf.Tensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
tf.Tensor
of shape(1,)
, optional, returned whenlabels
is provided) β Classification loss.logits (
tf.Tensor
of shape(batch_size, sequence_length, config.num_labels)
) β Classification scores (before SoftMax).mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β Tuple oftf.Tensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
TFXLNetForTokenClassificationOutput
ortuple(tf.Tensor)
Example:
>>> from transformers import XLNetTokenizer, TFXLNetForTokenClassification >>> import tensorflow as tf >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = TFXLNetForTokenClassification.from_pretrained('xlnet-base-cased') >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="tf") >>> input_ids = inputs["input_ids"] >>> inputs["labels"] = tf.reshape(tf.constant([1] * tf.size(input_ids).numpy()), (-1, tf.size(input_ids))) # Batch size 1 >>> outputs = model(inputs) >>> loss, scores = outputs[:2]
TFXLNetForQuestionAnsweringSimpleΒΆ
-
class
transformers.
TFXLNetForQuestionAnsweringSimple
(*args, **kwargs)[source]ΒΆ XLNet Model with a span classification head on top for extractive question-answering tasks like SQuAD (a linear layers on top of the hidden-states output to compute span start logits and span end logits).
Note
TF 2.0 models accepts two formats as inputs:
having all inputs as keyword arguments (like PyTorch models), or
having all inputs as a list, tuple or dict in the first positional arguments.
This second option is useful when using
tf.keras.Model.fit()
method which currently requires having all the tensors in the first argument of the model call function:model(inputs)
.If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :
a single Tensor with input_ids only and nothing else:
model(inputs_ids)
a list of varying length with one or several input Tensors IN THE ORDER given in the docstring:
model([input_ids, attention_mask])
ormodel([input_ids, attention_mask, token_type_ids])
a dictionary with one or several input Tensors associated to the input names given in the docstring:
model({'input_ids': input_ids, 'token_type_ids': token_type_ids})
- Parameters
config (
XLNetConfig
) β 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.
-
call
(inputs=None, attention_mask=None, mems=None, perm_mask=None, target_mapping=None, token_type_ids=None, input_mask=None, head_mask=None, inputs_embeds=None, use_cache=True, output_attentions=None, output_hidden_states=None, return_dict=None, start_positions=None, end_positions=None, training=False)[source]ΒΆ The
TFXLNetForQuestionAnsweringSimple
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.XLNetTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Mask values selected in
[0, 1]
:1
for tokens that are NOT MASKED,0
for MASKED tokens.mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input ids as they have already been computed.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) β Mask to indicate the attention pattern for each input token with values selected in[0, 1]
: Ifperm_mask[k, i, j] = 0
, i attend to j in batch k; ifperm_mask[k, i, j] = 1
, i does not attend to j in batch k. If None, each token attends to all the others (full bidirectional attention). Only used during pretraining (to define factorization order) or for sequential decoding (generation).target_mapping (
tf.Tensor
orNumpy array
of shape(batch_size, num_predict, sequence_length)
, optional) β Mask to indicate the output tokens to use. Iftarget_mapping[k, i, j] = 1
, the i-th predict in batch k is on the j-th token. Only used during pretraining for partial prediction or for sequential decoding (generation).token_type_ids (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βSegment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0
corresponds to a sentence A token,1
corresponds to a sentence B tokeninput_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) β Mask to avoid performing attention on padding token indices. Negative of attention_mask, i.e. with 0 for real tokens and 1 for padding. Kept for compatibility with the original code base. You can only uses one of input_mask and attention_mask Mask values selected in[0, 1]
:1
for tokens that are MASKED,0
for tokens that are NOT MASKED.head_mask (
tf.Tensor
orNumpy array
of shape(num_heads,)
or(num_layers, num_heads)
, optional) β 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 (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, hidden_size)
, optional) β Optionally, instead of passinginput_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 modelβs internal embedding lookup matrix.use_cache (
bool
) β If use_cache is True, mems are returned and can be used to speed up decoding (see mems). Defaults to True.output_attentions (
bool
, optional) β If set toTrue
, the attentions tensors of all attention layers are returned. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β If set toTrue
, the hidden states of all layers are returned. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β If set toTrue
, the model will return aModelOutput
instead of a plain tuple.start_positions (
tf.Tensor
of shape(batch_size,)
, optional) β Labels for position (index) of the start of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length). Position outside of the sequence are not taken into account for computing the loss.end_positions (
tf.Tensor
of shape(batch_size,)
, optional) β Labels for position (index) of the end of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length). Position outside of the sequence are not taken into account for computing the loss.
- Returns
A
TFXLNetForQuestionAnsweringSimpleOutput
(ifreturn_dict=True
is passed or whenconfig.return_dict=True
) or a tuple oftf.Tensor
comprising various elements depending on the configuration (XLNetConfig
) and inputs.loss (
tf.Tensor
of shape(1,)
, optional, returned whenlabels
is provided) β Total span extraction loss is the sum of a Cross-Entropy for the start and end positions.start_logits (
tf.Tensor
of shape(batch_size, sequence_length,)
) β Span-start scores (before SoftMax).end_logits (
tf.Tensor
of shape(batch_size, sequence_length,)
) β Span-end scores (before SoftMax).mems (
List[tf.Tensor]
of lengthconfig.n_layers
) β Contains pre-computed hidden-states. Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.hidden_states (
tuple(tf.Tensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) β Tuple oftf.Tensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(tf.Tensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) β Tuple oftf.Tensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
TFXLNetForQuestionAnsweringSimpleOutput
ortuple(tf.Tensor)
Example:
>>> from transformers import XLNetTokenizer, TFXLNetForQuestionAnsweringSimple >>> import tensorflow as tf >>> tokenizer = XLNetTokenizer.from_pretrained('xlnet-base-cased') >>> model = TFXLNetForQuestionAnsweringSimple.from_pretrained('xlnet-base-cased') >>> question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet" >>> input_dict = tokenizer(question, text, return_tensors='tf') >>> start_scores, end_scores = model(input_dict) >>> all_tokens = tokenizer.convert_ids_to_tokens(input_dict["input_ids"].numpy()[0]) >>> answer = ' '.join(all_tokens[tf.math.argmax(start_scores, 1)[0] : tf.math.argmax(end_scores, 1)[0]+1])