# Transformer XL¶

## Overview¶

The Transformer-XL model was proposed in Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context by Zihang Dai, Zhilin Yang, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov. It’s a causal (uni-directional) transformer with relative positioning (sinusoïdal) embeddings which can reuse previously computed hidden-states to attend to longer context (memory). This model also uses adaptive softmax inputs and outputs (tied).

The abstract from the paper is the following:

Transformers have a potential of learning longer-term dependency, but are limited by a fixed-length context in the setting of language modeling. We propose a novel neural architecture Transformer-XL that enables learning dependency beyond a fixed length without disrupting temporal coherence. It consists of a segment-level recurrence mechanism and a novel positional encoding scheme. Our method not only enables capturing longer-term dependency, but also resolves the context fragmentation problem. As a result, Transformer-XL learns dependency that is 80% longer than RNNs and 450% longer than vanilla Transformers, achieves better performance on both short and long sequences, and is up to 1,800+ times faster than vanilla Transformers during evaluation. Notably, we improve the state-of-the-art results of bpc/perplexity to 0.99 on enwiki8, 1.08 on text8, 18.3 on WikiText-103, 21.8 on One Billion Word, and 54.5 on Penn Treebank (without finetuning). When trained only on WikiText-103, Transformer-XL manages to generate reasonably coherent, novel text articles with thousands of tokens.

Tips:

• Transformer-XL uses relative sinusoidal positional embeddings. Padding can be done on the left or on the right. The original implementation trains on SQuAD with padding on the left, therefore the padding defaults are set to left.

• Transformer-XL 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.

## TransfoXLConfig¶

class transformers.TransfoXLConfig(vocab_size=267735, cutoffs=[20000, 40000, 200000], d_model=1024, d_embed=1024, n_head=16, d_head=64, d_inner=4096, div_val=4, pre_lnorm=False, n_layer=18, mem_len=1600, clamp_len=1000, same_length=True, proj_share_all_but_first=True, attn_type=0, sample_softmax=- 1, adaptive=True, dropout=0.1, dropatt=0.0, untie_r=True, init='normal', init_range=0.01, proj_init_std=0.01, init_std=0.02, layer_norm_epsilon=1e-05, eos_token_id=0, **kwargs)[source]

This is the configuration class to store the configuration of a TransfoXLModel or a TFTransfoXLModel. It is used to instantiate a Transformer-XL 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 Transformer XL 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 267735) – Vocabulary size of the BERT model. Defines the number of different tokens that can be represented by the inputs_ids passed when calling TransfoXLModel or TFTransfoXLModel.

• cutoffs (List[int], optional, defaults to [20000, 40000, 200000]) – Cutoffs for the adaptive softmax.

• d_model (int, optional, defaults to 1024) – Dimensionality of the model’s hidden states.

• d_embed (int, optional, defaults to 1024) – Dimensionality of the embeddings

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

• d_head (int, optional, defaults to 64) – Dimensionality of the model’s heads.

• d_inner (int, optional, defaults to 4096) – Inner dimension in FF

• div_val (int, optional, defaults to 4) – Divident value for adapative input and softmax

• pre_lnorm (boolean, optional, defaults to False) – Whether or not to apply LayerNorm to the input instead of the output in the blocks.

• n_layer (int, optional, defaults to 18) – Number of hidden layers in the Transformer encoder.

• mem_len (int, optional, defaults to 1600) – Length of the retained previous heads.

• clamp_len (int, optional, defaults to 1000) – Use the same pos embeddings after clamp_len.

• same_length (boolean, optional, defaults to True) – Whether or not to use the same attn length for all tokens

• proj_share_all_but_first (boolean, optional, defaults to True) – True to share all but first projs, False not to share.

• attn_type (int, optional, defaults to 0) – Attention type. 0 for Transformer-XL, 1 for Shaw et al, 2 for Vaswani et al, 3 for Al Rfou et al.

• sample_softmax (int, optional, defaults to -1) – Number of samples in the sampled softmax.

• adaptive (boolean, optional, defaults to True) – Whether or not to use adaptive softmax.

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

• dropatt (float, optional, defaults to 0) – The dropout ratio for the attention probabilities.

• untie_r (boolean, optional, defaults to True) – Whether ot not to untie relative position biases.

• init (str, optional, defaults to "normal") – Parameter initializer to use.

• init_range (float, optional, defaults to 0.01) – Parameters initialized by U(-init_range, init_range).

• proj_init_std (float, optional, defaults to 0.01) – Parameters initialized by N(0, init_std)

• init_std (float, optional, defaults to 0.02) – Parameters initialized by N(0, init_std)

• layer_norm_epsilon (float, optional, defaults to 1e-5) – The epsilon to use in the layer normalization layers

Examples:

>>> from transformers import TransfoXLConfig, TransfoXLModel

>>> # Initializing a Transformer XL configuration
>>> configuration = TransfoXLConfig()

>>> # Initializing a model from the configuration
>>> model = TransfoXLModel(configuration)

>>> # Accessing the model configuration
>>> configuration = model.config


## TransfoXLTokenizer¶

class transformers.TransfoXLTokenizer(special=None, min_freq=0, max_size=None, lower_case=False, delimiter=None, vocab_file=None, pretrained_vocab_file: str = None, never_split=None, unk_token='<unk>', eos_token='<eos>', additional_special_tokens=['<formula>'], language='en', **kwargs)[source]

Construct a Transformer-XL tokenizer adapted from Vocab class in the original code. The Transformer-XL tokenizer is a word-level tokenizer (no sub-word tokenization).

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
• special (List[str], optional) – A list of special tokens (to be treated by the original implementation of this tokenizer).

• min_freq (int, optional, defaults to 0) – The minimum number of times a token has to be present in order to be kept in the vocabulary (otherwise it will be mapped to unk_token).

• max_size (int, optional) – The maximum size of the vocabulary. If left unset, it will default to the size of the vocabulary found after excluding the tokens according to the min_freq rule.

• lower_case (bool, optional, defaults to False) – Whether or not to lowercase the input when tokenizing.

• delimiter (str, optional) – The delimiter used between tokens.

• vocab_file (str, optional) – File containing the vocabulary (from the original implementation).

• pretrained_vocab_file (str, optional) – File containing the vocabulary as saved with the save_pretrained() method.

• never_split (List[str], optional) – List of tokens that should never be split. If no list is specified, will simply use the existing special tokens.

• unk_token (str, optional, defaults to "<unk>") – The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this token instead.

• eos_token (str, optional, defaults to "<eos>") – The end of sequence token.

• additional_special_tokens (List[str], optional, defaults to ["<formula>"]) – A list of additional special tokens (for the HuggingFace functionality).

• language (str, optional, defaults to "en") – The language of this tokenizer (used for mose preprocessing).

save_vocabulary(save_directory: str, filename_prefix: Optional[str] = None) → Tuple[str][source]

Save only the vocabulary of the tokenizer (vocabulary + added tokens).

This method won’t 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) – The directory in which to save the vocabulary.

• filename_prefix (str, optional) – An optional prefix to add to the named of the saved files.

Returns

Paths to the files saved.

Return type

Tuple(str)

## TransfoXL specific outputs¶

class transformers.models.transfo_xl.modeling_transfo_xl.TransfoXLModelOutput(last_hidden_state: torch.FloatTensor, mems: List[torch.FloatTensor] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]

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

Parameters
• 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.

• mems (List[torch.FloatTensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) –

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

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

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

class transformers.models.transfo_xl.modeling_transfo_xl.TransfoXLLMHeadModelOutput(losses: Optional[torch.FloatTensor] = None, prediction_scores: torch.FloatTensor = None, mems: List[torch.FloatTensor] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]

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

Parameters
• losses (torch.FloatTensor of shape (batch_size, sequence_length-1), optional, returned when labels is provided) – Language modeling losses (not reduced).

• 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 after SoftMax).

• mems (List[torch.FloatTensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) –

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

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

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

class transformers.models.transfo_xl.modeling_tf_transfo_xl.TFTransfoXLModelOutput(last_hidden_state: tensorflow.python.framework.ops.Tensor = None, mems: 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]

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

Parameters
• last_hidden_state (tf.Tensor of shape (batch_size, sequence_length, hidden_size)) – Sequence of hidden-states at the output of the last layer of the model.

• mems (List[tf.Tensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(tf.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) –

Tuple 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 when output_attentions=True is passed or when config.output_attentions=True) –

Tuple 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.transfo_xl.modeling_tf_transfo_xl.TFTransfoXLLMHeadModelOutput(prediction_scores: tensorflow.python.framework.ops.Tensor = None, mems: 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]

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

Parameters
• losses (tf.Tensor of shape (batch_size, sequence_length-1), optional, returned when labels is provided) – Language modeling losses (not reduced).

• prediction_scores (tf.Tensor of shape (batch_size, sequence_length, config.vocab_size)) – Prediction scores of the language modeling head (scores for each vocabulary token after SoftMax).

• mems (List[tf.Tensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(tf.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) –

Tuple 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 when output_attentions=True is passed or when config.output_attentions=True) –

Tuple 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.

## TransfoXLModel¶

class transformers.TransfoXLModel(config)[source]

The bare Bert 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 (TransfoXLConfig) – 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, mems=None, head_mask=None, inputs_embeds=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]

The TransfoXLModel 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 TransfoXLTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

What are input IDs?

• mems (List[torch.FloatTensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input_ids as they have already been computed.

• head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional) –

Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:

• 1 indicates the head is not masked,

• 0 indicates the head is masked.

• inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

• output_attentions (bool, optional) – Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.

• output_hidden_states (bool, optional) – Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.

• return_dict (bool, optional) – Whether or not to return a ModelOutput instead of a plain tuple.

Returns

A TransfoXLModelOutput or a tuple of torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (TransfoXLConfig) and inputs.

• last_hidden_state (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size)) – Sequence of hidden-states at the output of the last layer of the model.

• mems (List[torch.FloatTensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) – Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, 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 output_attentions=True is passed or when config.output_attentions=True) – Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

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

Return type

TransfoXLModelOutput or tuple(torch.FloatTensor)

Example:

>>> from transformers import TransfoXLTokenizer, TransfoXLModel
>>> import torch

>>> tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
>>> model = TransfoXLModel.from_pretrained('transfo-xl-wt103')

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
>>> outputs = model(**inputs)

>>> last_hidden_states = outputs.last_hidden_state


## TransfoXLLMHeadModel¶

class transformers.TransfoXLLMHeadModel(config)[source]

The Transformer-XL Model with a language modeling head on top (adaptive softmax with weights tied to the adaptive 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 (TransfoXLConfig) – 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, mems=None, head_mask=None, inputs_embeds=None, labels=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]

The TransfoXLLMHeadModel 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 TransfoXLTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

What are input IDs?

• mems (List[torch.FloatTensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input_ids as they have already been computed.

• head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional) –

Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:

• 1 indicates the head is not masked,

• 0 indicates the head is masked.

• inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

• output_attentions (bool, optional) – Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.

• output_hidden_states (bool, optional) – Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.

• return_dict (bool, optional) – Whether or not to return a ModelOutput instead of a plain tuple.

• labels (torch.LongTensor of shape (batch_size, sequence_length), optional) – Labels for language modeling. Note that the labels are shifted inside the model, i.e. you can set 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

A TransfoXLLMHeadModelOutput or a tuple of torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (TransfoXLConfig) and inputs.

• losses (torch.FloatTensor of shape (batch_size, sequence_length-1), optional, returned when labels is provided) Language modeling losses (not reduced).

• 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 after SoftMax).

• mems (List[torch.FloatTensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) – Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, 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 output_attentions=True is passed or when config.output_attentions=True) – Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

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

Return type

TransfoXLLMHeadModelOutput or tuple(torch.FloatTensor)

Example:

>>> import torch
>>> from transformers import TransfoXLTokenizer, TransfoXLLMHeadModel

>>> tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
>>> model = TransfoXLLMHeadModel.from_pretrained('transfo-xl-wt103')

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
>>> outputs = model(**inputs, labels=inputs["input_ids"])
>>> loss = outputs.loss
>>> logits = outputs.logits


## TransfoXLForSequenceClassification¶

class transformers.TransfoXLForSequenceClassification(config)[source]

The Transformer-XL Model transformer with a sequence classification head on top (linear layer).

TransfoXLForSequenceClassification uses the last token in order to do the classification, as other causal models (e.g. GPT-1) do.

Since it does classification on the last token, it requires to know the position of the last token. If a pad_token_id is defined in the configuration, it finds the last token that is not a padding token in each row. If no pad_token_id is defined, it simply takes the last value in each row of the batch. Since it cannot guess the padding tokens when inputs_embeds are passed instead of input_ids, it does the same (take the last value in each row of the batch).

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 (TransfoXLConfig) – 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, mems=None, head_mask=None, inputs_embeds=None, labels=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]

The TransfoXLForSequenceClassification 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 TransfoXLTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

What are input IDs?

• mems (List[torch.FloatTensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input_ids as they have already been computed.

• head_mask (torch.FloatTensor of shape (num_heads,) or (num_layers, num_heads), optional) –

Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:

• 1 indicates the head is not masked,

• 0 indicates the head is masked.

• inputs_embeds (torch.FloatTensor of shape (batch_size, sequence_length, hidden_size), optional) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

• output_attentions (bool, optional) – Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail.

• output_hidden_states (bool, optional) – Whether or not to return the hidden states of all layers. See hidden_states under returned tensors for more detail.

• return_dict (bool, optional) – Whether or not to return a ModelOutput instead of a plain tuple.

• labels (torch.LongTensor of shape (batch_size,), optional) – 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

A TransfoXLSequenceClassifierOutputWithPast or a tuple of torch.FloatTensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (TransfoXLConfig) and inputs.

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

• mems (List[torch.FloatTensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(torch.FloatTensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) – Tuple of torch.FloatTensor (one for the output of the embeddings + one for the output of each layer) of shape (batch_size, 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 output_attentions=True is passed or when config.output_attentions=True) – Tuple of torch.FloatTensor (one for each layer) of shape (batch_size, num_heads, sequence_length, sequence_length).

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

Return type

TransfoXLSequenceClassifierOutputWithPast or tuple(torch.FloatTensor)

Example:

>>> from transformers import TransfoXLTokenizer, TransfoXLForSequenceClassification
>>> import torch

>>> tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
>>> model = TransfoXLForSequenceClassification.from_pretrained('transfo-xl-wt103')

>>> 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


## TFTransfoXLModel¶

class transformers.TFTransfoXLModel(*args, **kwargs)[source]

The bare Bert 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]) 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 (TransfoXLConfig) – 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(input_ids=None, mems=None, head_mask=None, inputs_embeds=None, output_attentions=None, output_hidden_states=None, return_dict=None, training=False, **kwargs)[source]

The TFTransfoXLModel 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 BertTokenizer. See transformers.PreTrainedTokenizer.__call__() and transformers.PreTrainedTokenizer.encode() for details.

What are input IDs?

• mems (List[tf.Tensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input_ids as they have already been computed.

• head_mask (tf.Tensor or Numpy array of shape (num_heads,) or (num_layers, num_heads), optional) –

Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:

• 1 indicates the head is not masked,

• 0 indicates the head is masked.

• inputs_embeds (tf.Tensor or Numpy array of shape (batch_size, sequence_length, hidden_size), optional) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

• output_attentions (bool, optional) – Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail. 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) – Whether or not to return the hidden states of all layers. See hidden_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) – Whether or not to return a ModelOutput 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 to False) – Whether or not to use the model in training mode (some modules like dropout modules have different behaviors between training and evaluation).

Returns

A TFTransfoXLModelOutput or a tuple of tf.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (TransfoXLConfig) and inputs.

• last_hidden_state (tf.Tensor of shape (batch_size, sequence_length, hidden_size)) – Sequence of hidden-states at the output of the last layer of the model.

• mems (List[tf.Tensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(tf.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) – Tuple 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 when output_attentions=True is passed or when config.output_attentions=True) – Tuple 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.

Return type

TFTransfoXLModelOutput or tuple(tf.Tensor)

Example:

>>> from transformers import TransfoXLTokenizer, TFTransfoXLModel
>>> import tensorflow as tf

>>> tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
>>> model = TFTransfoXLModel.from_pretrained('transfo-xl-wt103')

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="tf")
>>> outputs = model(inputs)

>>> last_hidden_states = outputs.last_hidden_state


## TFTransfoXLLMHeadModel¶

class transformers.TFTransfoXLLMHeadModel(*args, **kwargs)[source]

The Transformer-XL Model with a language modeling head on top (adaptive softmax with weights tied to the adaptive 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]) 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 (TransfoXLConfig) – 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(input_ids=None, mems=None, head_mask=None, inputs_embeds=None, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False, **kwargs)[source]

The TFTransfoXLLMHeadModel 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 BertTokenizer. See transformers.PreTrainedTokenizer.__call__() and transformers.PreTrainedTokenizer.encode() for details.

What are input IDs?

• mems (List[tf.Tensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input_ids as they have already been computed.

• head_mask (tf.Tensor or Numpy array of shape (num_heads,) or (num_layers, num_heads), optional) –

Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:

• 1 indicates the head is not masked,

• 0 indicates the head is masked.

• inputs_embeds (tf.Tensor or Numpy array of shape (batch_size, sequence_length, hidden_size), optional) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

• output_attentions (bool, optional) – Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail. 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) – Whether or not to return the hidden states of all layers. See hidden_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) – Whether or not to return a ModelOutput 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 to False) – Whether or not to use the model in training mode (some modules like dropout modules have different behaviors between training and evaluation).

Returns

A TFTransfoXLLMHeadModelOutput or a tuple of tf.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (TransfoXLConfig) and inputs.

• losses (tf.Tensor of shape (batch_size, sequence_length-1), optional, returned when labels is provided) Language modeling losses (not reduced).

• prediction_scores (tf.Tensor of shape (batch_size, sequence_length, config.vocab_size)) – Prediction scores of the language modeling head (scores for each vocabulary token after SoftMax).

• mems (List[tf.Tensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(tf.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) – Tuple 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 when output_attentions=True is passed or when config.output_attentions=True) – Tuple 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.

Return type

TFTransfoXLLMHeadModelOutput or tuple(tf.Tensor)

Example:

>>> from transformers import TransfoXLTokenizer, TFTransfoXLLMHeadModel
>>> import tensorflow as tf

>>> tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
>>> model = TFTransfoXLLMHeadModel.from_pretrained('transfo-xl-wt103')

>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="tf")
>>> outputs = model(inputs)
>>> logits = outputs.logits


## TFTransfoXLForSequenceClassification¶

class transformers.TFTransfoXLForSequenceClassification(*args, **kwargs)[source]

The Transfo XL Model transformer with a sequence classification head on top (linear layer).

TFTransfoXLForSequenceClassification uses the last token in order to do the classification, as other causal models (e.g. GPT-1,GPT-2) do.

Since it does classification on the last token, it requires to know the position of the last token. If a pad_token_id is defined in the configuration, it finds the last token that is not a padding token in each row. If no pad_token_id is defined, it simply takes the last value in each row of the batch. Since it cannot guess the padding tokens when inputs_embeds are passed instead of input_ids, it does the same (take the last value in each row of the batch).

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]) 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 (TransfoXLConfig) – 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(input_ids=None, mems=None, head_mask=None, inputs_embeds=None, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False, **kwargs)[source]

The TFTransfoXLForSequenceClassification 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 BertTokenizer. See transformers.PreTrainedTokenizer.__call__() and transformers.PreTrainedTokenizer.encode() for details.

What are input IDs?

• mems (List[tf.Tensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks) as computed by the model (see mems output below). Can be used to speed up sequential decoding. The token ids which have their mems given to this model should not be passed as input_ids as they have already been computed.

• head_mask (tf.Tensor or Numpy array of shape (num_heads,) or (num_layers, num_heads), optional) –

Mask to nullify selected heads of the self-attention modules. Mask values selected in [0, 1]:

• 1 indicates the head is not masked,

• 0 indicates the head is masked.

• inputs_embeds (tf.Tensor or Numpy array of shape (batch_size, sequence_length, hidden_size), optional) – Optionally, instead of passing input_ids you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

• output_attentions (bool, optional) – Whether or not to return the attentions tensors of all attention layers. See attentions under returned tensors for more detail. 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) – Whether or not to return the hidden states of all layers. See hidden_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) – Whether or not to return a ModelOutput 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 to False) – 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) – Labels for computing the cross entropy classification loss. Indices should be in [0, ..., config.vocab_size - 1].

Returns

A TFTransfoXLSequenceClassifierOutputWithPast or a tuple of tf.Tensor (if return_dict=False is passed or when config.return_dict=False) comprising various elements depending on the configuration (TransfoXLConfig) and inputs.

• loss (tf.Tensor of shape (1,), optional, returned when labels is provided) – Classification (or regression if config.num_labels==1) loss.

• logits (tf.Tensor of shape (batch_size, config.num_labels)) – Classification (or regression if config.num_labels==1) scores (before SoftMax).

• mems (List[tf.Tensor] of length config.n_layers) – Contains pre-computed hidden-states (key and values in the attention blocks). Can be used (see mems input) to speed up sequential decoding. The token ids which have their past given to this model should not be passed as input ids as they have already been computed.

• hidden_states (tuple(tf.Tensor), optional, returned when output_hidden_states=True is passed or when config.output_hidden_states=True) – Tuple 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 when output_attentions=True is passed or when config.output_attentions=True) – Tuple 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.

Return type

TFTransfoXLSequenceClassifierOutputWithPast or tuple(tf.Tensor)

Example:

>>> from transformers import TransfoXLTokenizer, TFTransfoXLForSequenceClassification
>>> import tensorflow as tf

>>> tokenizer = TransfoXLTokenizer.from_pretrained('transfo-xl-wt103')
>>> model = TFTransfoXLForSequenceClassification.from_pretrained('transfo-xl-wt103')

>>> 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


## Internal Layers¶

class transformers.AdaptiveEmbedding(n_token, d_embed, d_proj, cutoffs, div_val=1, sample_softmax=False)[source]
class transformers.TFAdaptiveEmbedding(*args, **kwargs)[source]