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
andtarget_mapping
inputs to control the attention span and outputs (see examples in examples/pytorch/text-generation/run_generation.py)XLNet is one of the few models that has no sequence length limit.
This model was contributed by thomwolf. 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=512, reuse_len=None, use_mems_eval=True, use_mems_train=False, 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
or aTFXLNetModel
. It is used to instantiate a 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 number of different tokens that can be represented by theinputs_ids
passed when callingXLNetModel
orTFXLNetModel
.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 (often named feed-forward) layer in the Transformer encoder.ff_activation (
str
orCallable
, optional, defaults to"gelu"
) β The non-linear activation function (function or string) in the If string,"gelu"
,"relu"
,"silu"
and"gelu_new"
are supported.untie_r (
bool
, optional, defaults toTrue
) β Whether or not to untie relative position biasesattn_type (
str
, optional, defaults to"bi"
) β The attention type used by the model. Set"bi"
for XLNet,"uni"
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
, optional) β The number of tokens in the current batch to be cached and reused in the future.bi_data (
bool
, optional, defaults toFalse
) β Whether or not to use bidirectional input pipeline. Usually set toTrue
during pretraining andFalse
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 (
bool
, optional, defaults toFalse
) β Whether or not to use the same attention length for each token.summary_type (
str
, optional, defaults to βlastβ) βArgument used when doing sequence summary. Used in the sequence classification and multiple choice models.
Has to be one of the following options:
"last"
: Take the last token hidden state (like XLNet)."first"
: Take the first token hidden state (like BERT)."mean"
: Take the mean of all tokens hidden states."cls_index"
: Supply a Tensor of classification token position (like GPT/GPT-2)."attn"
: Not implemented now, use multi-head attention.
summary_use_proj (
bool
, optional, defaults toTrue
) βArgument used when doing sequence summary. Used in the sequence classification and multiple choice models.
Whether or not to add a projection after the vector extraction.
summary_activation (
str
, optional) βArgument used when doing sequence summary. Used in the sequence classification and multiple choice models.
Pass
"tanh"
for a tanh activation to the output, any other value will result in no activation.summary_proj_to_labels (
boo
, optional, defaults toTrue
) βUsed in the sequence classification and multiple choice models.
Whether the projection outputs should have
config.num_labels
orconfig.hidden_size
classes.summary_last_dropout (
float
, optional, defaults to 0.1) βUsed in the sequence classification and multiple choice models.
The dropout ratio to be used after the projection and activation.
start_n_top (
int
, optional, defaults to 5) β Used in the SQuAD evaluation script.end_n_top (
int
, optional, defaults to 5) β Used in the SQuAD evaluation script.use_mems_eval (
bool
, optional, defaults toTrue
) β Whether or not the model should make use of the recurrent memory mechanism in evaluation mode.use_mems_train (
bool
, optional, defaults toFalse
) βWhether or not the model should make use of the recurrent memory mechanism in train mode.
Note
For pretraining, it is recommended to set
use_mems_train
toTrue
. For fine-tuning, it is recommended to setuse_mems_train
toFalse
as discussed here. Ifuse_mems_train
is set toTrue
, one has to make sure that the train batches are correctly pre-processed, e.g.batch_1 = [[This line is], [This is the]]
andbatch_2 = [[ the first line], [ second line]]
and that all batches are of equal size.
Examples:
>>> 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>'], sp_model_kwargs: Optional[Dict[str, Any]] = None, **kwargs)[source]ΒΆ Construct an XLNet tokenizer. Based on SentencePiece.
This tokenizer inherits from
PreTrainedTokenizer
which contains most of the main methods. Users should refer to this superclass for more information regarding those methods.- Parameters
vocab_file (
str
) 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 (
str
, optional, defaults to"<s>"
) βThe beginning of sequence token that was used during pretraining. 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 (
str
, optional, defaults to"</s>"
) 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 (
str
, optional, defaults to"<unk>"
) 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 (
str
, optional, defaults to"<sep>"
) 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 (
str
, optional, defaults to"<pad>"
) β The token used for padding, for example when batching sequences of different lengths.cls_token (
str
, optional, defaults to"<cls>"
) 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 (
str
, optional, defaults to"<mask>"
) 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_kwargs (
dict
, optional) βWill be passed to the
SentencePieceProcessor.__init__()
method. The Python wrapper for SentencePiece can be used, among other things, to set:enable_sampling
: Enable subword regularization.nbest_size
: Sampling parameters for unigram. Invalid for BPE-Dropout.nbest_size = {0,1}
: No sampling is performed.nbest_size > 1
: samples from the nbest_size results.nbest_size < 0
: assuming that nbest_size is infinite and samples from the all hypothesis (lattice) using forward-filtering-and-backward-sampling algorithm.
alpha
: Smoothing parameter for unigram sampling, and dropout probability of merge operations for BPE-dropout.
-
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 added.token_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 Create 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 0 1 1 1 1 1 1 1 1 1 | first sequence | second sequence |
If
token_ids_1
isNone
, this method only returns the first portion of the mask (0s).- 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 Retrieve sequence ids from a token list that has no special tokens added. This method is called when adding special tokens using the tokenizer
prepare_for_model
method.- 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 Whether or not the token list is already formatted with special tokens for the model.
- Returns
A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
- Return type
List[int]
-
save_vocabulary
(save_directory: str, filename_prefix: Optional[str] = None) → Tuple[str][source]ΒΆ Save only the vocabulary of the tokenizer (vocabulary + added tokens).
This method wonβt save the configuration and special token mappings of the tokenizer. Use
_save_pretrained()
to save the whole state of the tokenizer.- Parameters
save_directory (
str
) β The directory in which to save the vocabulary.filename_prefix (
str
, optional) β An optional prefix to add to the named of the saved files.
- Returns
Paths to the files saved.
- Return type
Tuple(str)
XLNetTokenizerFastΒΆ
-
class
transformers.
XLNetTokenizerFast
(vocab_file=None, tokenizer_file=None, 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 Construct a βfastβ XLNet tokenizer (backed by HuggingFaceβs tokenizers library). Based on Unigram.
This tokenizer inherits from
PreTrainedTokenizerFast
which contains most of the main methods. Users should refer to this superclass for more information regarding those methods.- Parameters
vocab_file (
str
) 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 (
str
, optional, defaults to"<s>"
) βThe beginning of sequence token that was used during pretraining. 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 (
str
, optional, defaults to"</s>"
) 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 (
str
, optional, defaults to"<unk>"
) 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 (
str
, optional, defaults to"<sep>"
) 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 (
str
, optional, defaults to"<pad>"
) β The token used for padding, for example when batching sequences of different lengths.cls_token (
str
, optional, defaults to"<cls>"
) 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 (
str
, optional, defaults to"<mask>"
) 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 added.token_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 Create 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 0 1 1 1 1 1 1 1 1 1 | first sequence | second sequence |
If
token_ids_1
isNone
, this method only returns the first portion of the mask (0s).- 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]
-
save_vocabulary
(save_directory: str, filename_prefix: Optional[str] = None) → Tuple[str][source]ΒΆ Save only the vocabulary of the tokenizer (vocabulary + added tokens).
This method wonβt save the configuration and special token mappings of the tokenizer. Use
_save_pretrained()
to save the whole state of the tokenizer.- Parameters
save_directory (
str
) β The directory in which to save the vocabulary.filename_prefix (
str
, optional) β An optional prefix to add to the named of the saved files.
- Returns
Paths to the files saved.
- Return type
Tuple(str)
-
slow_tokenizer_class
ΒΆ alias of
transformers.models.xlnet.tokenization_xlnet.XLNetTokenizer
XLNet specific outputsΒΆ
-
class
transformers.models.xlnet.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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_xlnet.
XLNetLMHeadModelOutput
(loss: Optional[torch.FloatTensor] = None, 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
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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_xlnet.
XLNetForSequenceClassificationOutput
(loss: Optional[torch.FloatTensor] = None, 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
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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_xlnet.
XLNetForMultipleChoiceOutput
(loss: Optional[torch.FloatTensor] = None, 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
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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_xlnet.
XLNetForTokenClassificationOutput
(loss: Optional[torch.FloatTensor] = None, 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
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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_xlnet.
XLNetForQuestionAnsweringSimpleOutput
(loss: Optional[torch.FloatTensor] = None, start_logits: Optional[torch.FloatTensor] = None, end_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
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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_tf_xlnet.
TFXLNetModelOutput
(last_hidden_state: Optional[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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_tf_xlnet.
TFXLNetLMHeadModelOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: Optional[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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_tf_xlnet.
TFXLNetForSequenceClassificationOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: Optional[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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_tf_xlnet.
TFXLNetForMultipleChoiceOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: Optional[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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_tf_xlnet.
TFXLNetForTokenClassificationOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: Optional[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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.models.xlnet.modeling_tf_xlnet.
TFXLNetForQuestionAnsweringSimpleOutput
(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, start_logits: Optional[tensorflow.python.framework.ops.Tensor] = None, end_logits: Optional[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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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 inherits from
PreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- 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, use_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, **kwargs)[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 shapebatch_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 (
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.
input_mask (
torch.FloatTensor
of shapebatch_size, sequence_length
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β Whether or not to return aModelOutput
instead of a plain tuple.
- Returns
A
XLNetModelOutput
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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') >>> 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 inherits from
PreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- 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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, **kwargs)[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 shapebatch_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 (
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.
input_mask (
torch.FloatTensor
of shapebatch_size, sequence_length
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β Whether or not to return aModelOutput
instead of a plain tuple.labels (
torch.LongTensor
of shape(batch_size, num_predict)
, optional) βLabels for masked language modeling.
num_predict
corresponds totarget_mapping.shape[1]
. Iftarget_mapping
is :obj`None`, thennum_predict
corresponds tosequence_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 theinput_ids
(see theprepare_inputs_for_generation
function 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]
- Returns
A
XLNetLMHeadModelOutput
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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') >>> # 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 inherits from
PreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- 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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, **kwargs)[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 shapebatch_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 (
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.
input_mask (
torch.FloatTensor
of shapebatch_size, sequence_length
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β Whether or not to 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
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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') >>> 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 inherits from
PreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- 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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, **kwargs)[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 shapebatch_size, num_choices, 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 (
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.
input_mask (
torch.FloatTensor
of shapebatch_size, num_choices, sequence_length
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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, num_choices, 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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β Whether or not to 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-1]
wherenum_choices
is the size of the second dimension of the input tensors. (Seeinput_ids
above)
- Returns
A
XLNetForMultipleChoiceOutput
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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') >>> 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 inherits from
PreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- 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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, **kwargs)[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 shapebatch_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 (
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.
input_mask (
torch.FloatTensor
of shapebatch_size, sequence_length
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β Whether or not to 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
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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') >>> 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 inherits from
PreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- 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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, **kwargs)[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 shapebatch_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 (
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.
input_mask (
torch.FloatTensor
of shapebatch_size, sequence_length
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β Whether or not to 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
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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') >>> 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 inherits from
PreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a PyTorch torch.nn.Module subclass. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- 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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, **kwargs)[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 shapebatch_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 (
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
torch.FloatTensor
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.
input_mask (
torch.FloatTensor
of shapebatch_size, sequence_length
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) β Whether or not to 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.is_impossible (
torch.LongTensor
of shape(batch_size,)
, optional) β Labels whether a question has an answer or no answer (SQuAD 2.0)cls_index (
torch.LongTensor
of shape(batch_size,)
, optional) β 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) β 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.
- Returns
A
XLNetForQuestionAnsweringOutput
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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') >>> 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 outputting raw hidden-states without any specific head on top.
This model inherits from
TFPreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a tf.keras.Model subclass. Use it as a regular TF 2.0 Keras Model and refer to the TF 2.0 documentation for all matter related to general usage and behavior.
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
(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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, training=False, **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 (
Numpy array
ortf.Tensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
BertTokenizer
. Seetransformers.PreTrainedTokenizer.__call__()
andtransformers.PreTrainedTokenizer.encode()
for details.attention_mask (
Numpy array
ortf.Tensor
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.:obj:
use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.token_type_ids (
Numpy array
ortf.Tensor
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.
input_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
Numpy array
ortf.Tensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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
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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.return_dict (
bool
, optional) β Whether or not to return aModelOutput
instead of a plain tuple. This argument can be used in eager mode, in graph mode the value will always be set to True.training (
bool
, optional, defaults toFalse
) β Whether or not to use the model in training mode (some modules like dropout modules have different behaviors between training and evaluation).
- Returns
A
TFXLNetModelOutput
or a tuple oftf.Tensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.last_hidden_state
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).
This model inherits from
TFPreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a tf.keras.Model subclass. Use it as a regular TF 2.0 Keras Model and refer to the TF 2.0 documentation for all matter related to general usage and behavior.
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
(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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False, **kwargs)[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 (
Numpy array
ortf.Tensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
BertTokenizer
. Seetransformers.PreTrainedTokenizer.__call__()
andtransformers.PreTrainedTokenizer.encode()
for details.attention_mask (
Numpy array
ortf.Tensor
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.:obj:
use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.token_type_ids (
Numpy array
ortf.Tensor
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.
input_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
Numpy array
ortf.Tensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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
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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.return_dict (
bool
, optional) β Whether or not to return aModelOutput
instead of a plain tuple. This argument can be used in eager mode, in graph mode the value will always be set to True.training (
bool
, optional, defaults toFalse
) β Whether or not to use the model in training mode (some modules like dropout modules have different behaviors between training and evaluation).labels (
tf.Tensor
of shape(batch_size, sequence_length)
, optional) β Labels for computing the cross entropy classification loss. Indices should be in[0, ..., config.vocab_size - 1]
.
- Returns
A
TFXLNetLMHeadModelOutput
or a tuple oftf.Tensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.
This model inherits from
TFPreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a tf.keras.Model subclass. Use it as a regular TF 2.0 Keras Model and refer to the TF 2.0 documentation for all matter related to general usage and behavior.
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
(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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False, **kwargs)[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 (
Numpy array
ortf.Tensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
BertTokenizer
. Seetransformers.PreTrainedTokenizer.__call__()
andtransformers.PreTrainedTokenizer.encode()
for details.attention_mask (
Numpy array
ortf.Tensor
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.:obj:
use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.token_type_ids (
Numpy array
ortf.Tensor
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.
input_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
Numpy array
ortf.Tensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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
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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.return_dict (
bool
, optional) β Whether or not to return aModelOutput
instead of a plain tuple. This argument can be used in eager mode, in graph mode the value will always be set to True.training (
bool
, optional, defaults toFalse
) β Whether or not to use the model in training mode (some modules like dropout modules have different behaviors between training and evaluation).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
or a tuple oftf.Tensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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 = outputs.loss >>> logits = outputs.logits
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.
This model inherits from
TFPreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a tf.keras.Model subclass. Use it as a regular TF 2.0 Keras Model and refer to the TF 2.0 documentation for all matter related to general usage and behavior.
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
(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, use_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False, **kwargs)[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 (
Numpy array
ortf.Tensor
of shape(batch_size, num_choices, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
BertTokenizer
. Seetransformers.PreTrainedTokenizer.__call__()
andtransformers.PreTrainedTokenizer.encode()
for details.attention_mask (
Numpy array
ortf.Tensor
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.:obj:
use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.token_type_ids (
Numpy array
ortf.Tensor
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.
input_mask (
tf.Tensor
orNumpy array
of shape(batch_size, num_choices, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
Numpy array
ortf.Tensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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
of shape(batch_size, num_choices, 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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.return_dict (
bool
, optional) β Whether or not to return aModelOutput
instead of a plain tuple. This argument can be used in eager mode, in graph mode the value will always be set to True.training (
bool
, optional, defaults toFalse
) β Whether or not to use the model in training mode (some modules like dropout modules have different behaviors between training and evaluation).labels (
tf.Tensor
of shape(batch_size,)
, optional) β Labels for computing the multiple choice classification loss. Indices should be in[0, ..., num_choices]
wherenum_choices
is the size of the second dimension of the input tensors. (Seeinput_ids
above)
- Returns
A
TFXLNetForMultipleChoiceOutput
or a tuple oftf.Tensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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.logits
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.
This model inherits from
TFPreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a tf.keras.Model subclass. Use it as a regular TF 2.0 Keras Model and refer to the TF 2.0 documentation for all matter related to general usage and behavior.
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
(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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False, **kwargs)[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 (
Numpy array
ortf.Tensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
BertTokenizer
. Seetransformers.PreTrainedTokenizer.__call__()
andtransformers.PreTrainedTokenizer.encode()
for details.attention_mask (
Numpy array
ortf.Tensor
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.:obj:
use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.token_type_ids (
Numpy array
ortf.Tensor
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.
input_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
Numpy array
ortf.Tensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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
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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.return_dict (
bool
, optional) β Whether or not to return aModelOutput
instead of a plain tuple. This argument can be used in eager mode, in graph mode the value will always be set to True.training (
bool
, optional, defaults toFalse
) β Whether or not to use the model in training mode (some modules like dropout modules have different behaviors between training and evaluation).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
or a tuple oftf.Tensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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 = outputs.loss >>> logits = outputs.logits
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).
This model inherits from
TFPreTrainedModel
. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)This model is also a tf.keras.Model subclass. Use it as a regular TF 2.0 Keras Model and refer to the TF 2.0 documentation for all matter related to general usage and behavior.
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
(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_mems=None, output_attentions=None, output_hidden_states=None, return_dict=None, start_positions=None, end_positions=None, training=False, **kwargs)[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 (
Numpy array
ortf.Tensor
of shape(batch_size, sequence_length)
) βIndices of input sequence tokens in the vocabulary.
Indices can be obtained using
BertTokenizer
. Seetransformers.PreTrainedTokenizer.__call__()
andtransformers.PreTrainedTokenizer.encode()
for details.attention_mask (
Numpy array
ortf.Tensor
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 tokens that are masked.
mems (
List[torch.FloatTensor]
of lengthconfig.n_layers
) βContains pre-computed hidden-states (see
mems
output below) . Can be used to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_ids
as they have already been computed.:obj:
use_mems
has to be set toTrue
to make use ofmems
.perm_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length, sequence_length)
, optional) βMask to indicate the attention pattern for each input token with values selected in
[0, 1]
:if
perm_mask[k, i, j] = 0
, i attend to j in batch k;if
perm_mask[k, i, j] = 1
, i does not attend to j in batch k.
If not set, 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.token_type_ids (
Numpy array
ortf.Tensor
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.
input_mask (
tf.Tensor
orNumpy array
of shape(batch_size, sequence_length)
, optional) βMask to avoid performing attention on padding token indices. Negative of
attention_mask
, i.e. with 0 for real tokens and 1 for padding which is kept for compatibility with the original code base.Mask values selected in
[0, 1]
:1 for tokens that are masked,
0 for tokens that are not masked.
You can only uses one of
input_mask
andattention_mask
.head_mask (
Numpy array
ortf.Tensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) βMask 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
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 convertinput_ids
indices into associated vectors than the modelβs internal embedding lookup matrix.output_attentions (
bool
, optional) β Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.output_hidden_states (
bool
, optional) β Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail. This argument can be used only in eager mode, in graph mode the value in the config will be used instead.return_dict (
bool
, optional) β Whether or not to return aModelOutput
instead of a plain tuple. This argument can be used in eager mode, in graph mode the value will always be set to True.training (
bool
, optional, defaults toFalse
) β Whether or not to use the model in training mode (some modules like dropout modules have different behaviors between training and evaluation).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
or a tuple oftf.Tensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) 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 (seemems
input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed asinput_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') >>> outputs = model(input_dict) >>> start_logits = outputs.start_logits >>> end_logits = outputs.end_logits >>> all_tokens = tokenizer.convert_ids_to_tokens(input_dict["input_ids"].numpy()[0]) >>> answer = ' '.join(all_tokens[tf.math.argmax(start_logits, 1)[0] : tf.math.argmax(end_logits, 1)[0]+1])