XLM

Overview

The XLM model was proposed in Cross-lingual Language Model Pretraining by Guillaume Lample*, Alexis Conneau*. It’s a transformer pre-trained using one of the following objectives:

  • a causal language modeling (CLM) objective (next token prediction),

  • a masked language modeling (MLM) objective (Bert-like), or

  • a Translation Language Modeling (TLM) object (extension of Bert’s MLM to multiple language inputs)

The abstract from the paper is the following:

Recent studies have demonstrated the efficiency of generative pretraining for English natural language understanding. In this work, we extend this approach to multiple languages and show the effectiveness of cross-lingual pretraining. We propose two methods to learn cross-lingual language models (XLMs): one unsupervised that only relies on monolingual data, and one supervised that leverages parallel data with a new cross-lingual language model objective. We obtain state-of-the-art results on cross-lingual classification, unsupervised and supervised machine translation. On XNLI, our approach pushes the state of the art by an absolute gain of 4.9% accuracy. On unsupervised machine translation, we obtain 34.3 BLEU on WMT’16 German-English, improving the previous state of the art by more than 9 BLEU. On supervised machine translation, we obtain a new state of the art of 38.5 BLEU on WMT’16 Romanian-English, outperforming the previous best approach by more than 4 BLEU. Our code and pretrained models will be made publicly available.

Tips:

  • XLM has many different checkpoints, which were trained using different objectives: CLM, MLM or TLM. Make sure to select the correct objective for your task (e.g. MLM checkpoints are not suitable for generation).

  • XLM has multilingual checkpoints which leverage a specific lang parameter. Check out the multi-lingual page for more information.

XLMConfig

class transformers.XLMConfig(vocab_size=30145, emb_dim=2048, n_layers=12, n_heads=16, dropout=0.1, attention_dropout=0.1, gelu_activation=True, sinusoidal_embeddings=False, causal=False, asm=False, n_langs=1, use_lang_emb=True, max_position_embeddings=512, embed_init_std=0.02209708691207961, layer_norm_eps=1e-12, init_std=0.02, bos_index=0, eos_index=1, pad_index=2, unk_index=3, mask_index=5, is_encoder=True, summary_type='first', summary_use_proj=True, summary_activation=None, summary_proj_to_labels=True, summary_first_dropout=0.1, start_n_top=5, end_n_top=5, mask_token_id=0, lang_id=0, **kwargs)[source]

This is the configuration class to store the configuration of a XLMModel. It is used to instantiate an XLM 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 xlm-mlm-en-2048 architecture.

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

Parameters
  • vocab_size (int, optional, defaults to 30145) – Vocabulary size of the XLM model. Defines the different tokens that can be represented by the inputs_ids passed to the forward method of XLMModel.

  • emb_dim (int, optional, defaults to 2048) – Dimensionality of the encoder layers and the pooler layer.

  • n_layer (int, optional, defaults to 12) – Number of hidden layers in the Transformer encoder.

  • n_head (int, optional, defaults to 16) – Number of attention heads for each attention layer in the Transformer encoder.

  • dropout (float, optional, defaults to 0.1) – The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.

  • attention_dropout (float, optional, defaults to 0.1) – The dropout probability for the attention mechanism

  • gelu_activation (boolean, optional, defaults to True) – The non-linear activation function (function or string) in the encoder and pooler. If set to True, “gelu” will be used instead of “relu”.

  • sinusoidal_embeddings (boolean, optional, defaults to False) – Whether to use sinusoidal positional embeddings instead of absolute positional embeddings.

  • causal (boolean, optional, defaults to False) – Set this to True for the model to behave in a causal manner. Causal models use a triangular attention mask in order to only attend to the left-side context instead if a bidirectional context.

  • asm (boolean, optional, defaults to False) – Whether to use an adaptive log softmax projection layer instead of a linear layer for the prediction layer.

  • n_langs (int, optional, defaults to 1) – The number of languages the model handles. Set to 1 for monolingual models.

  • use_lang_emb (boolean, optional, defaults to True) – Whether to use language embeddings. Some models use additional language embeddings, see the multilingual models page for information on how to use them.

  • max_position_embeddings (int, optional, defaults to 512) – The maximum sequence length that this model might ever be used with. Typically set this to something large just in case (e.g., 512 or 1024 or 2048).

  • embed_init_std (float, optional, defaults to 2048^-0.5) – The standard deviation of the truncated_normal_initializer for initializing the embedding matrices.

  • init_std (int, optional, defaults to 50257) – The standard deviation of the truncated_normal_initializer for initializing all weight matrices except the embedding matrices.

  • layer_norm_eps (float, optional, defaults to 1e-12) – The epsilon used by the layer normalization layers.

  • bos_index (int, optional, defaults to 0) – The index of the beginning of sentence token in the vocabulary.

  • eos_index (int, optional, defaults to 1) – The index of the end of sentence token in the vocabulary.

  • pad_index (int, optional, defaults to 2) – The index of the padding token in the vocabulary.

  • unk_index (int, optional, defaults to 3) – The index of the unknown token in the vocabulary.

  • mask_index (int, optional, defaults to 5) – The index of the masking token in the vocabulary.

  • is_encoder (boolean, optional, defaults to True) – Whether the initialized model should be a transformer encoder or decoder as seen in Vaswani et al.

  • summary_type (string, optional, defaults to “first”) –

    Argument used when doing sequence summary. Used in for the multiple choice head in XLMForSequenceClassification. Is 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 (GPT/GPT-2)

    • ’attn’ => Not implemented now, use multi-head attention

  • summary_use_proj (boolean, optional, defaults to True) – Argument used when doing sequence summary. Used in for the multiple choice head in XLMForSequenceClassification. Add a projection after the vector extraction

  • summary_activation (string or None, optional, defaults to None) – Argument used when doing sequence summary. Used in for the multiple choice head in XLMForSequenceClassification. ‘tanh’ => add a tanh activation to the output, Other => no activation.

  • summary_proj_to_labels (boolean, optional, defaults to True) – Argument used when doing sequence summary. Used in for the multiple choice head in XLMForSequenceClassification. If True, the projection outputs to config.num_labels classes (otherwise to hidden_size). Default: False.

  • summary_first_dropout (float, optional, defaults to 0.1) – Argument used when doing sequence summary. Used in for the multiple choice head in XLMForSequenceClassification. Add a dropout before the projection and activation

  • start_n_top (int, optional, defaults to 5) – Used in the SQuAD evaluation script for XLM and XLNet.

  • end_n_top (int, optional, defaults to 5) – Used in the SQuAD evaluation script for XLM and XLNet.

  • mask_token_id (int, optional, defaults to 0) – Model agnostic parameter to identify masked tokens when generating text in an MLM context.

  • lang_id (int, optional, defaults to 1) – The ID of the language used by the model. This parameter is used when generating text in a given language.

Example:

from transformers import XLMConfig, XLMModel

# Initializing a XLM configuration
configuration = XLMConfig()

# Initializing a model from the configuration
model = XLMModel(configuration)

# Accessing the model configuration
configuration = model.config
pretrained_config_archive_map

A dictionary containing all the available pre-trained checkpoints.

Type

Dict[str, str]

XLMTokenizer

class transformers.XLMTokenizer(vocab_file, merges_file, unk_token='<unk>', bos_token='<s>', sep_token='</s>', pad_token='<pad>', cls_token='</s>', mask_token='<special1>', additional_special_tokens=['<special0>', '<special1>', '<special2>', '<special3>', '<special4>', '<special5>', '<special6>', '<special7>', '<special8>', '<special9>'], lang2id=None, id2lang=None, do_lowercase_and_remove_accent=True, **kwargs)[source]

BPE tokenizer for XLM

  • Moses preprocessing & tokenization for most supported languages

  • Language specific tokenization for Chinese (Jieba), Japanese (KyTea) and Thai (PyThaiNLP)

  • (optionally) lower case & normalize all inputs text

  • argument special_tokens and function set_special_tokens, can be used to add additional symbols (ex: “__classify__”) to a vocabulary

  • lang2id attribute maps the languages supported by the model with their ids if provided (automatically set for pretrained vocabularies)

  • id2lang attributes does reverse mapping if provided (automatically set for pretrained vocabularies)

  • do_lowercase_and_remove_accent controle lower casing and accent (automatically set for pretrained vocabularies)

build_inputs_with_special_tokens(token_ids_0, token_ids_1=None)[source]

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

single sequence: <s> X </s> pair of sequences: <s> A </s> B </s>

convert_tokens_to_string(tokens)[source]

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

create_token_type_ids_from_sequences(token_ids_0, token_ids_1=None)[source]

Creates a mask from the two sequences passed to be used in a sequence-pair classification task. An XLM sequence pair mask has the following format: 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 | first sequence | second sequence

if token_ids_1 is None, only returns the first portion of the mask (0’s).

get_special_tokens_mask(token_ids_0, token_ids_1=None, already_has_special_tokens=False)[source]

Retrieves sequence ids from a token list that has no special tokens added. This method is called when adding special tokens using the tokenizer prepare_for_model or encode_plus methods.

Parameters
  • token_ids_0 – list of ids (must not contain special tokens)

  • token_ids_1 – Optional list of ids (must not contain special tokens), necessary when fetching sequence ids for sequence pairs

  • already_has_special_tokens – (default False) Set to True if the token list is already formated with special tokens for the model

Returns

1 for a special token, 0 for a sequence token.

Return type

A list of integers in the range [0, 1]

save_vocabulary(save_directory)[source]

Save the tokenizer vocabulary and merge files to a directory.

property vocab_size

Size of the base vocabulary (without the added tokens)

XLMModel

class transformers.XLMModel(config)[source]

The bare XLM Model transformer outputting raw hidden-states without any specific head on top.

This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

Parameters

config (XLMConfig) – 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 the from_pretrained() method to load the model weights.

forward(input_ids=None, attention_mask=None, langs=None, token_type_ids=None, position_ids=None, lengths=None, cache=None, head_mask=None, inputs_embeds=None)[source]

The XLMModel forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (torch.LongTensor of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (torch.LongTensor of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, torch.FloatTensor], optional, defaults to None) – dictionary with torch.FloatTensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

Returns

last_hidden_state (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size)):

Sequence of hidden-states at the output of the last layer of the model.

hidden_states (tuple(torch.FloatTensor), optional, returned when config.output_hidden_states=True):

Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, 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 when config.output_attentions=True):

Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

tuple(torch.FloatTensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

from transformers import XLMTokenizer, XLMModel
import torch

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = XLMModel.from_pretrained('xlm-mlm-en-2048')
input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True)).unsqueeze(0)  # Batch size 1
outputs = model(input_ids)
last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
get_input_embeddings()[source]

Returns the model’s input embeddings.

Returns

A torch module mapping vocabulary to hidden states.

Return type

nn.Module

set_input_embeddings(new_embeddings)[source]

Set model’s input embeddings

Parameters

value (nn.Module) – A module mapping vocabulary to hidden states.

XLMWithLMHeadModel

class transformers.XLMWithLMHeadModel(config)[source]

The XLM Model transformer with a language modeling head on top (linear layer with weights tied to the input embeddings).

This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

Parameters

config (XLMConfig) – 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 the from_pretrained() method to load the model weights.

forward(input_ids=None, attention_mask=None, langs=None, token_type_ids=None, position_ids=None, lengths=None, cache=None, head_mask=None, inputs_embeds=None, labels=None)[source]

The XLMWithLMHeadModel forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (torch.LongTensor of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (torch.LongTensor of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, torch.FloatTensor], optional, defaults to None) – dictionary with torch.FloatTensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  • labels (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) – Labels for language modeling. Note that the labels are shifted inside the model, i.e. you can set lm_labels = input_ids Indices are selected in [-100, 0, ..., config.vocab_size] All labels set to -100 are ignored (masked), the loss is only computed for labels in [0, ..., config.vocab_size]

Returns

loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided)

Language modeling loss.

prediction_scores (torch.FloatTensor of shape (batch_size, sequence_length, config.vocab_size)):

Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).

hidden_states (tuple(torch.FloatTensor), optional, returned when config.output_hidden_states=True):

Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, 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 when config.output_attentions=True):

Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

tuple(torch.FloatTensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

from transformers import XLMTokenizer, XLMWithLMHeadModel
import torch

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = XLMWithLMHeadModel.from_pretrained('xlm-mlm-en-2048')
input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True)).unsqueeze(0)  # Batch size 1
outputs = model(input_ids)
last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
get_output_embeddings()[source]

Returns the model’s output embeddings.

Returns

A torch module mapping hidden states to vocabulary.

Return type

nn.Module

XLMForSequenceClassification

class transformers.XLMForSequenceClassification(config)[source]

XLM Model with a sequence classification/regression head on top (a linear layer on top of the pooled output) e.g. for GLUE tasks.

This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

Parameters

config (XLMConfig) – 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 the from_pretrained() method to load the model weights.

forward(input_ids=None, attention_mask=None, langs=None, token_type_ids=None, position_ids=None, lengths=None, cache=None, head_mask=None, inputs_embeds=None, labels=None)[source]

The XLMForSequenceClassification forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (torch.LongTensor of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (torch.LongTensor of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, torch.FloatTensor], optional, defaults to None) – dictionary with torch.FloatTensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  • labels (torch.LongTensor of shape (batch_size,), optional, defaults to None) – Labels for computing the sequence classification/regression loss. Indices should be in [0, ..., config.num_labels - 1]. If config.num_labels == 1 a regression loss is computed (Mean-Square loss), If config.num_labels > 1 a classification loss is computed (Cross-Entropy).

Returns

loss (torch.FloatTensor of shape (1,), optional, returned when label is provided):

Classification (or regression if config.num_labels==1) loss.

logits (torch.FloatTensor of shape (batch_size, config.num_labels)):

Classification (or regression if config.num_labels==1) scores (before SoftMax).

hidden_states (tuple(torch.FloatTensor), optional, returned when config.output_hidden_states=True):

Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, 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 when config.output_attentions=True):

Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

tuple(torch.FloatTensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

from transformers import XLMTokenizer, XLMForSequenceClassification
import torch

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = XLMForSequenceClassification.from_pretrained('xlm-mlm-en-2048')
input_ids = torch.tensor(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True)).unsqueeze(0)  # Batch size 1
labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
outputs = model(input_ids, labels=labels)
loss, logits = outputs[:2]

XLMForQuestionAnsweringSimple

class transformers.XLMForQuestionAnsweringSimple(config)[source]

XLM Model with a span classification head on top for extractive question-answering tasks like SQuAD (a linear layers on top of the hidden-states output to compute span start logits and span end logits).

This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

Parameters

config (XLMConfig) – 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 the from_pretrained() method to load the model weights.

forward(input_ids=None, attention_mask=None, langs=None, token_type_ids=None, position_ids=None, lengths=None, cache=None, head_mask=None, inputs_embeds=None, start_positions=None, end_positions=None)[source]

The XLMForQuestionAnsweringSimple forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (torch.LongTensor of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (torch.LongTensor of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, torch.FloatTensor], optional, defaults to None) – dictionary with torch.FloatTensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  • start_positions (torch.LongTensor of shape (batch_size,), optional, defaults to None) – 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, defaults to None) – 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

loss (torch.FloatTensor of shape (1,), optional, returned when labels is provided):

Total span extraction loss is the sum of a Cross-Entropy for the start and end positions.

start_scores (torch.FloatTensor of shape (batch_size, sequence_length,)):

Span-start scores (before SoftMax).

end_scores (torch.FloatTensor of shape (batch_size, sequence_length,)):

Span-end scores (before SoftMax).

hidden_states (tuple(torch.FloatTensor), optional, returned when config.output_hidden_states=True):

Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, 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 when config.output_attentions=True):

Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

tuple(torch.FloatTensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

from transformers import XLMTokenizer, XLMForQuestionAnsweringSimple
import torch

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = XLMForQuestionAnsweringSimple.from_pretrained('xlm-mlm-en-2048')
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[0]

XLMForQuestionAnswering

class transformers.XLMForQuestionAnswering(config)[source]

XLM Model with a beam-search span classification head on top for extractive question-answering tasks like SQuAD (a linear layers on top of the hidden-states output to compute span start logits and span end logits).

This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.

Parameters

config (XLMConfig) – 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 the from_pretrained() method to load the model weights.

forward(input_ids=None, attention_mask=None, langs=None, token_type_ids=None, position_ids=None, lengths=None, cache=None, head_mask=None, inputs_embeds=None, start_positions=None, end_positions=None, is_impossible=None, cls_index=None, p_mask=None)[source]

The XLMForQuestionAnswering forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (torch.LongTensor of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (torch.LongTensor of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (torch.LongTensor of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, torch.FloatTensor], optional, defaults to None) – dictionary with torch.FloatTensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  • start_positions (torch.LongTensor of shape (batch_size,), optional, defaults to None) – 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, defaults to None) – 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, defaults to None) – Labels whether a question has an answer or no answer (SQuAD 2.0)

  • cls_index (torch.LongTensor of shape (batch_size,), optional, defaults to None) – 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, defaults to None) – Optional mask of tokens which can’t be in answers (e.g. [CLS], [PAD], …). 1.0 means token should be masked. 0.0 mean token is not masked.

Returns

loss (torch.FloatTensor of shape (1,), optional, returned if both start_positions and end_positions are provided):

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 if start_positions or end_positions is not provided):

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 if start_positions or end_positions is not provided):

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 if start_positions or end_positions is not provided):

Log probabilities for the top config.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 if start_positions or end_positions is not provided):

Indices for the top config.start_n_top * config.end_n_top end token possibilities (beam-search).

cls_logits (torch.FloatTensor of shape (batch_size,), optional, returned if start_positions or end_positions is not provided):

Log probabilities for the is_impossible label of the answers.

hidden_states (tuple(torch.FloatTensor), optional, returned when config.output_hidden_states=True):

Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, 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 when config.output_attentions=True):

Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

tuple(torch.FloatTensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

from transformers import XLMTokenizer, XLMForQuestionAnswering
import torch

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = XLMForQuestionAnswering.from_pretrained('xlm-mlm-en-2048')
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[0]

TFXLMModel

class transformers.TFXLMModel(config, *inputs, **kwargs)[source]

The bare XLM Model transformer outputing raw hidden-states without any specific head on top.

Note

TF 2.0 models accepts two formats as inputs:

  • having all inputs as keyword arguments (like PyTorch models), or

  • having all inputs as a list, tuple or dict in the first positional arguments.

This second option is useful when using tf.keras.Model.fit() method which currently requires having all the tensors in the first argument of the model call function: model(inputs).

If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :

  • a single Tensor with input_ids only and nothing else: model(inputs_ids)

  • a list of varying length with one or several input Tensors IN THE ORDER given in the docstring: model([input_ids, attention_mask]) or model([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 (XLMConfig) – 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 the from_pretrained() method to load the model weights.

call(inputs, **kwargs)[source]

The TFXLMModel forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (tf.Tensor or Numpy array of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, tf.Tensor], optional, defaults to None) – dictionary with tf.Tensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (tf.Tensor or Numpy array of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (tf.Tensor or Numpy array of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

Returns

last_hidden_state (tf.Tensor or Numpy array of shape (batch_size, sequence_length, hidden_size)):

Sequence of hidden-states at the output of the last layer of the model.

hidden_states (tuple(tf.Tensor), optional, returned when config.output_hidden_states=True):

Tuple of tf.Tensor or Numpy array (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 when config.output_attentions=True):

Tuple of tf.Tensor or Numpy array (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

tuple(tf.Tensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

import tensorflow as tf
from transformers import XLMTokenizer, TFXLMModel

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = TFXLMModel.from_pretrained('xlm-mlm-en-2048')
input_ids = tf.constant(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True))[None, :]  # Batch size 1
outputs = model(input_ids)
last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple

TFXLMWithLMHeadModel

class transformers.TFXLMWithLMHeadModel(config, *inputs, **kwargs)[source]

The XLM Model transformer with a language modeling head on top (linear layer with weights tied to the input embeddings).

Note

TF 2.0 models accepts two formats as inputs:

  • having all inputs as keyword arguments (like PyTorch models), or

  • having all inputs as a list, tuple or dict in the first positional arguments.

This second option is useful when using tf.keras.Model.fit() method which currently requires having all the tensors in the first argument of the model call function: model(inputs).

If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :

  • a single Tensor with input_ids only and nothing else: model(inputs_ids)

  • a list of varying length with one or several input Tensors IN THE ORDER given in the docstring: model([input_ids, attention_mask]) or model([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 (XLMConfig) – 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 the from_pretrained() method to load the model weights.

call(inputs, **kwargs)[source]

The TFXLMWithLMHeadModel forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (tf.Tensor or Numpy array of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, tf.Tensor], optional, defaults to None) – dictionary with tf.Tensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (tf.Tensor or Numpy array of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (tf.Tensor or Numpy array of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

Returns

prediction_scores (tf.Tensor or Numpy array of shape (batch_size, sequence_length, config.vocab_size)):

Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).

hidden_states (tuple(tf.Tensor), optional, returned when config.output_hidden_states=True):

Tuple of tf.Tensor or Numpy array (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 when config.output_attentions=True):

Tuple of tf.Tensor or Numpy array (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

tuple(tf.Tensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

import tensorflow as tf
from transformers import XLMTokenizer, TFXLMWithLMHeadModel

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = TFXLMWithLMHeadModel.from_pretrained('xlm-mlm-en-2048')
input_ids = tf.constant(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True))[None, :]  # Batch size 1
outputs = model(input_ids)
last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output tuple
get_output_embeddings()[source]

Returns the model’s output embeddings.

Returns

A torch module mapping hidden states to vocabulary.

Return type

tf.keras.layers.Layer

TFXLMForSequenceClassification

class transformers.TFXLMForSequenceClassification(config, *inputs, **kwargs)[source]

XLM Model with a sequence classification/regression head on top (a linear layer on top of the pooled output) e.g. for GLUE tasks.

Note

TF 2.0 models accepts two formats as inputs:

  • having all inputs as keyword arguments (like PyTorch models), or

  • having all inputs as a list, tuple or dict in the first positional arguments.

This second option is useful when using tf.keras.Model.fit() method which currently requires having all the tensors in the first argument of the model call function: model(inputs).

If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :

  • a single Tensor with input_ids only and nothing else: model(inputs_ids)

  • a list of varying length with one or several input Tensors IN THE ORDER given in the docstring: model([input_ids, attention_mask]) or model([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 (XLMConfig) – 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 the from_pretrained() method to load the model weights.

call(inputs, **kwargs)[source]

The TFXLMForSequenceClassification forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (tf.Tensor or Numpy array of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, tf.Tensor], optional, defaults to None) – dictionary with tf.Tensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (tf.Tensor or Numpy array of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (tf.Tensor or Numpy array of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

Returns

logits (tf.Tensor or Numpy array of shape (batch_size, config.num_labels)):

Classification (or regression if config.num_labels==1) scores (before SoftMax).

hidden_states (tuple(tf.Tensor), optional, returned when config.output_hidden_states=True):

Tuple of tf.Tensor or Numpy array (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 when config.output_attentions=True):

Tuple of tf.Tensor or Numpy array (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

tuple(tf.Tensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

import tensorflow as tf
from transformers import XLMTokenizer, TFXLMForSequenceClassification

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = TFXLMForSequenceClassification.from_pretrained('xlm-mlm-en-2048')
input_ids = tf.constant(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True))[None, :]  # Batch size 1
labels = tf.constant([1])[None, :]  # Batch size 1
outputs = model(input_ids)
logits = outputs[0]

TFXLMForQuestionAnsweringSimple

class transformers.TFXLMForQuestionAnsweringSimple(config, *inputs, **kwargs)[source]

XLM Model with a span classification head on top for extractive question-answering tasks like SQuAD (a linear layers on top of the hidden-states output to compute span start logits and span end logits).

Note

TF 2.0 models accepts two formats as inputs:

  • having all inputs as keyword arguments (like PyTorch models), or

  • having all inputs as a list, tuple or dict in the first positional arguments.

This second option is useful when using tf.keras.Model.fit() method which currently requires having all the tensors in the first argument of the model call function: model(inputs).

If you choose this second option, there are three possibilities you can use to gather all the input Tensors in the first positional argument :

  • a single Tensor with input_ids only and nothing else: model(inputs_ids)

  • a list of varying length with one or several input Tensors IN THE ORDER given in the docstring: model([input_ids, attention_mask]) or model([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 (XLMConfig) – 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 the from_pretrained() method to load the model weights.

call(inputs, **kwargs)[source]

The TFXLMForQuestionAnsweringSimple forward method, overrides the __call__() special method.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.

Parameters
  • input_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length)) –

    Indices of input sequence tokens in the vocabulary.

    Indices can be obtained using transformers.BertTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.encode_plus() for details.

    What are input IDs?

  • attention_mask (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]: 1 for tokens that are NOT MASKED, 0 for MASKED tokens.

    What are attention masks?

  • langs (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    A parallel sequence of tokens to be used to indicate the language of each token in the input. Indices are languages ids which can be obtained from the language names by using two conversion mappings provided in the configuration of the model (only provided for multilingual models). More precisely, the language name -> language id mapping is in model.config.lang2id (dict str -> int) and the language id -> language name mapping is model.config.id2lang (dict int -> str).

    See usage examples detailed in the multilingual documentation.

  • token_type_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Segment 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

    What are token type IDs?

  • position_ids (tf.Tensor or Numpy array of shape (batch_size, sequence_length), optional, defaults to None) –

    Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, config.max_position_embeddings - 1].

    What are position IDs?

  • lengths (tf.Tensor or Numpy array of shape (batch_size,), optional, defaults to None) – Length of each sentence that can be used to avoid performing attention on padding token indices. You can also use attention_mask for the same result (see above), kept here for compatbility. Indices selected in [0, ..., input_ids.size(-1)]:

  • cache (Dict[str, tf.Tensor], optional, defaults to None) – dictionary with tf.Tensor that contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see cache output below). Can be used to speed up sequential decoding. The dictionary object will be modified in-place during the forward pass to add newly computed hidden-states.

  • head_mask (tf.Tensor or Numpy array of shape (num_heads,) or (num_layers, num_heads), optional, defaults to None) – Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]: 1 indicates the head is not masked, 0 indicates the head is masked.

  • input_embeds (tf.Tensor or Numpy array of shape (batch_size, sequence_length, hidden_size), optional, defaults to None) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

Returns

start_scores (tf.Tensor or Numpy array of shape (batch_size, sequence_length,)):

Span-start scores (before SoftMax).

end_scores (tf.Tensor or Numpy array of shape (batch_size, sequence_length,)):

Span-end scores (before SoftMax).

hidden_states (tuple(tf.Tensor), optional, returned when config.output_hidden_states=True):

Tuple of tf.Tensor or Numpy array (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 when config.output_attentions=True):

Tuple of tf.Tensor or Numpy array (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

tuple(tf.Tensor) comprising various elements depending on the configuration (XLMConfig) and inputs

Examples:

import tensorflow as tf
from transformers import XLMTokenizer, TFXLMForQuestionAnsweringSimple

tokenizer = XLMTokenizer.from_pretrained('xlm-mlm-en-2048')
model = TFXLMForQuestionAnsweringSimple.from_pretrained('xlm-mlm-en-2048')
input_ids = tf.constant(tokenizer.encode("Hello, my dog is cute", add_special_tokens=True))[None, :]  # Batch size 1
outputs = model(input_ids)
start_scores, end_scores = outputs[:2]