LED

Overview

The LED model was proposed in Longformer: The Long-Document Transformer by Iz Beltagy, Matthew E. Peters, Arman Cohan.

The abstract from the paper is the following:

Transformer-based models are unable to process long sequences due to their self-attention operation, which scales quadratically with the sequence length. To address this limitation, we introduce the Longformer with an attention mechanism that scales linearly with sequence length, making it easy to process documents of thousands of tokens or longer. Longformer’s attention mechanism is a drop-in replacement for the standard self-attention and combines a local windowed attention with a task motivated global attention. Following prior work on long-sequence transformers, we evaluate Longformer on character-level language modeling and achieve state-of-the-art results on text8 and enwik8. In contrast to most prior work, we also pretrain Longformer and finetune it on a variety of downstream tasks. Our pretrained Longformer consistently outperforms RoBERTa on long document tasks and sets new state-of-the-art results on WikiHop and TriviaQA. We finally introduce the Longformer-Encoder-Decoder (LED), a Longformer variant for supporting long document generative sequence-to-sequence tasks, and demonstrate its effectiveness on the arXiv summarization dataset.

Tips:

• LEDForConditionalGeneration is an extension of BartForConditionalGeneration exchanging the traditional self-attention layer with Longformer’s chunked self-attention layer. LEDTokenizer is an alias of BartTokenizer.

• LED works very well on long-range sequence-to-sequence tasks where the input_ids largely exceed a length of 1024 tokens.

• LED pads the input_ids to be a multiple of config.attention_window if required. Therefore a small speed-up is gained, when LEDTokenizer is used with the pad_to_multiple_of argument.

• LED makes use of global attention by means of the global_attention_mask (see LongformerModel). For summarization, it is advised to put global attention only on the first <s> token. For question answering, it is advised to put global attention on all tokens of the question.

• To fine-tune LED on all 16384, it is necessary to enable gradient checkpointing by setting config.gradient_checkpointing = True.

• A notebook showing how to evaluate LED, can be accessed here.

• A notebook showing how to fine-tune LED, can be accessed here.

This model was contributed by patrickvonplaten.

LEDConfig

class transformers.LEDConfig(vocab_size=50265, max_encoder_position_embeddings=16384, max_decoder_position_embeddings=1024, encoder_layers=12, encoder_ffn_dim=4096, encoder_attention_heads=16, decoder_layers=12, decoder_ffn_dim=4096, decoder_attention_heads=16, encoder_layerdrop=0.0, decoder_layerdrop=0.0, use_cache=True, is_encoder_decoder=True, activation_function='gelu', d_model=1024, dropout=0.1, attention_dropout=0.0, activation_dropout=0.0, init_std=0.02, decoder_start_token_id=2, classifier_dropout=0.0, pad_token_id=1, bos_token_id=0, eos_token_id=2, gradient_checkpointing=False, attention_window: Union[List[int], int] = 512, **kwargs)

This is the configuration class to store the configuration of a LEDModel. It is used to instantiate an LED 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 LED allenai/led-base-16384 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 50265) – Vocabulary size of the LED model. Defines the number of different tokens that can be represented by the inputs_ids passed when calling LEDModel or TFLEDModel.

• d_model (int, optional, defaults to 1024) – Dimensionality of the layers and the pooler layer.

• encoder_layers (int, optional, defaults to 12) – Number of encoder layers.

• decoder_layers (int, optional, defaults to 12) – Number of decoder layers.

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

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

• decoder_ffn_dim (int, optional, defaults to 4096) – Dimensionality of the “intermediate” (often named feed-forward) layer in decoder.

• encoder_ffn_dim (int, optional, defaults to 4096) – Dimensionality of the “intermediate” (often named feed-forward) layer in decoder.

• activation_function (str or function, optional, defaults to "gelu") – The non-linear activation function (function or string) in the encoder and pooler. If string, "gelu", "relu", "silu" and "gelu_new" are supported.

• 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 0.0) – The dropout ratio for the attention probabilities.

• activation_dropout (float, optional, defaults to 0.0) – The dropout ratio for activations inside the fully connected layer.

• classifier_dropout (float, optional, defaults to 0.0) – The dropout ratio for classifier.

• max_encoder_position_embeddings (int, optional, defaults to 16384) – The maximum sequence length that the encoder might ever be used with.

• max_decoder_position_embeddings (int, optional, defaults to 16384) – The maximum sequence length that the decoder might ever be used with.

• init_std (float, optional, defaults to 0.02) – The standard deviation of the truncated_normal_initializer for initializing all weight matrices.

• encoder_layerdrop – (float, optional, defaults to 0.0): The LayerDrop probability for the encoder. See the LayerDrop paper for more details.

• decoder_layerdrop – (float, optional, defaults to 0.0): The LayerDrop probability for the decoder. See the LayerDrop paper for more details.

• use_cache (bool, optional, defaults to True) – Whether or not the model should return the last key/values attentions (not used by all models)

• gradient_checkpointing (bool, optional, defaults to False) – If True, use gradient checkpointing to save memory at the expense of slower backward pass.

• Example:: –

• from transformers import LEDModel (>>>) –

• LEDConfig –

• # Initializing a LED allenai/led-base-16384 style configuration (>>>) –

• configuration = LEDConfig() (>>>) –

• # Initializing a model from the allenai/led-base-16384 style configuration (>>>) –

• model = LEDModel (>>>) –

• # Accessing the model configuration (>>>) –

• configuration = model.config (>>>) –

LEDTokenizer

class transformers.LEDTokenizer(vocab_file, merges_file, errors='replace', bos_token='<s>', eos_token='</s>', sep_token='</s>', cls_token='<s>', unk_token='<unk>', pad_token='<pad>', mask_token='<mask>', add_prefix_space=False, **kwargs)

Construct a LED tokenizer.

LEDTokenizer is identical to BartTokenizer and runs end-to-end tokenization: punctuation splitting and wordpiece.

Refer to superclass BartTokenizer for usage examples and documentation concerning parameters.

build_inputs_with_special_tokens(token_ids_0: List[int], token_ids_1: Optional[List[int]] = None) → List[int]

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

• single sequence: <s> X </s>

• pair of sequences: <s> A </s></s> B </s>

Parameters

• token_ids_0 (List[int]) – List of IDs to which the special tokens will be added.

• token_ids_1 (List[int], optional) – Optional second list of IDs for sequence pairs.

Returns

List of input IDs with the appropriate special tokens.

Return type

List[int]

create_token_type_ids_from_sequences(token_ids_0: List[int], token_ids_1: Optional[List[int]] = None) → List[int]

Create a mask from the two sequences passed to be used in a sequence-pair classification task. RoBERTa does not make use of token type ids, therefore a list of zeros is returned.

Parameters

• token_ids_0 (List[int]) – List of IDs.

• token_ids_1 (List[int], optional) – Optional second list of IDs for sequence pairs.

Returns

List of zeros.

Return type

List[int]

get_special_tokens_mask(token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False) → List[int]

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

Parameters

• token_ids_0 (List[int]) – List of IDs.

• token_ids_1 (List[int], optional) – Optional second list of IDs for sequence pairs.

• already_has_special_tokens (bool, optional, defaults to False) – Whether or not the token list is already formatted with special tokens for the model.

Returns

A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.

Return type

List[int]

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

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)

LEDTokenizerFast

class transformers.LEDTokenizerFast(vocab_file, merges_file, tokenizer_file=None, errors='replace', bos_token='<s>', eos_token='</s>', sep_token='</s>', cls_token='<s>', unk_token='<unk>', pad_token='<pad>', mask_token='<mask>', add_prefix_space=False, **kwargs)

Construct a “fast” LED tokenizer (backed by HuggingFace’s tokenizers library).

LEDTokenizerFast is identical to BartTokenizerFast and runs end-to-end tokenization: punctuation splitting and wordpiece.

Refer to superclass BartTokenizerFast for usage examples and documentation concerning parameters.

slow_tokenizer_class

alias of transformers.models.led.tokenization_led.LEDTokenizer

LED specific outputs

class transformers.models.led.modeling_led.LEDEncoderBaseModelOutput(last_hidden_state: torch.FloatTensor, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None, global_attentions: Optional[Tuple[torch.FloatTensor]] = None)

Base class for LEDEncoder’s outputs, with potential hidden states, local and global attentions.

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.

• 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, x + attention_window + 1), where x is the number of tokens with global attention mask.

  Local attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token in the sequence to every token with global attention (first x values) and to every token in the attention window (remaining attention_window + 1 values). Note that the first x values refer to tokens with fixed positions in the text, but the remaining attention_window + 1 values refer to tokens with relative positions: the attention weight of a token to itself is located at index x + attention_window / 2 and the attention_window / 2 preceding (succeeding) values are the attention weights to the attention_window / 2 preceding (succeeding) tokens. If the attention window contains a token with global attention, the attention weight at the corresponding index is set to 0; the value should be accessed from the first x attention weights. If a token has global attention, the attention weights to all other tokens in attentions is set to 0, the values should be accessed from global_attentions.

• global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

class transformers.models.led.modeling_led.LEDSeq2SeqModelOutput(last_hidden_state: torch.FloatTensor = None, past_key_values: Optional[List[torch.FloatTensor]] = None, decoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None, decoder_attentions: Optional[Tuple[torch.FloatTensor]] = None, cross_attentions: Optional[Tuple[torch.FloatTensor]] = None, encoder_last_hidden_state: Optional[torch.FloatTensor] = None, encoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None, encoder_attentions: Optional[Tuple[torch.FloatTensor]] = None, encoder_global_attentions: Optional[Tuple[torch.FloatTensor]] = None)

Base class for model encoder’s outputs that also contains : pre-computed hidden states that can 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 decoder of the model.

  If past_key_values is used only the last hidden-state of the sequences of shape (batch_size, 1, hidden_size) is output.

• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) –

  List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_heads, sequence_length, embed_size_per_head)).

  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(torch.FloatTensor), optional, returned 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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• encoder_global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

class transformers.models.led.modeling_led.LEDSeq2SeqLMOutput(loss: Optional[torch.FloatTensor] = None, logits: torch.FloatTensor = None, past_key_values: Optional[List[torch.FloatTensor]] = None, decoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None, decoder_attentions: Optional[Tuple[torch.FloatTensor]] = None, cross_attentions: Optional[Tuple[torch.FloatTensor]] = None, encoder_last_hidden_state: Optional[torch.FloatTensor] = None, encoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None, encoder_attentions: Optional[Tuple[torch.FloatTensor]] = None, encoder_global_attentions: Optional[Tuple[torch.FloatTensor]] = None)

Base class for sequence-to-sequence language models outputs.

Parameters

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

• logits (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).

• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) –

  List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_heads, sequence_length, embed_size_per_head)).

  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(torch.FloatTensor), optional, returned 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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• encoder_global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

class transformers.models.led.modeling_led.LEDSeq2SeqSequenceClassifierOutput(loss: Optional[torch.FloatTensor] = None, logits: torch.FloatTensor = None, past_key_values: Optional[List[torch.FloatTensor]] = None, decoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None, decoder_attentions: Optional[Tuple[torch.FloatTensor]] = None, cross_attentions: Optional[Tuple[torch.FloatTensor]] = None, encoder_last_hidden_state: Optional[torch.FloatTensor] = None, encoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None, encoder_attentions: Optional[Tuple[torch.FloatTensor]] = None, encoder_global_attentions: Optional[Tuple[torch.FloatTensor]] = None)

Base class for outputs of sequence-to-sequence sentence classification models.

Parameters

• 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).

• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) –

  List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_heads, sequence_length, embed_size_per_head)).

  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(torch.FloatTensor), optional, returned 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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• encoder_global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

class transformers.models.led.modeling_led.LEDSeq2SeqQuestionAnsweringModelOutput(loss: Optional[torch.FloatTensor] = None, start_logits: torch.FloatTensor = None, end_logits: torch.FloatTensor = None, past_key_values: Optional[List[torch.FloatTensor]] = None, decoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None, decoder_attentions: Optional[Tuple[torch.FloatTensor]] = None, cross_attentions: Optional[Tuple[torch.FloatTensor]] = None, encoder_last_hidden_state: Optional[torch.FloatTensor] = None, encoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None, encoder_attentions: Optional[Tuple[torch.FloatTensor]] = None, encoder_global_attentions: Optional[Tuple[torch.FloatTensor]] = None)

Base class for outputs of sequence-to-sequence question answering models.

Parameters

• 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_logits (torch.FloatTensor of shape (batch_size, sequence_length)) – Span-start scores (before SoftMax).

• end_logits (torch.FloatTensor of shape (batch_size, sequence_length)) – Span-end scores (before SoftMax).

• past_key_values (List[torch.FloatTensor], optional, returned when use_cache=True is passed or when config.use_cache=True) –

  List of torch.FloatTensor of length config.n_layers, with each tensor of shape (2, batch_size, num_heads, sequence_length, embed_size_per_head)).

  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(torch.FloatTensor), optional, returned 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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• encoder_global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

class transformers.models.led.modeling_tf_led.TFLEDEncoderBaseModelOutput(last_hidden_state: 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, global_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)

Base class for Longformer’s outputs, with potential hidden states, local and global attentions.

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.

• 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, x + attention_window + 1), where x is the number of tokens with global attention mask.

  Local attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token in the sequence to every token with global attention (first x values) and to every token in the attention window (remaining attention_window + 1 values). Note that the first x values refer to tokens with fixed positions in the text, but the remaining attention_window + 1 values refer to tokens with relative positions: the attention weight of a token to itself is located at index x + attention_window / 2 and the attention_window / 2 preceding (succeeding) values are the attention weights to the attention_window / 2 preceding (succeeding) tokens. If the attention window contains a token with global attention, the attention weight at the corresponding index is set to 0; the value should be accessed from the first x attention weights. If a token has global attention, the attention weights to all other tokens in attentions is set to 0, the values should be accessed from global_attentions.

• global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

class transformers.models.led.modeling_tf_led.TFLEDSeq2SeqModelOutput(last_hidden_state: tensorflow.python.framework.ops.Tensor = None, past_key_values: Optional[List[tensorflow.python.framework.ops.Tensor]] = None, decoder_hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, decoder_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, cross_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, encoder_last_hidden_state: Optional[tensorflow.python.framework.ops.Tensor] = None, encoder_hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, encoder_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, encoder_global_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)

Base class for model encoder’s outputs that also contains : pre-computed hidden states that can 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 decoder of the model.

  If past_key_values is used only the last hidden-state of the sequences of shape (batch_size, 1, hidden_size) is output.

• past_key_values (List[tf.Tensor], optional, returned when use_cache=True is passed or when config.use_cache=True) –

  List of tf.Tensor of length config.n_layers, with each tensor of shape (2, batch_size, num_heads, sequence_length, embed_size_per_head)).

  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• encoder_global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

class transformers.models.led.modeling_tf_led.TFLEDSeq2SeqLMOutput(loss: Optional[tensorflow.python.framework.ops.Tensor] = None, logits: tensorflow.python.framework.ops.Tensor = None, past_key_values: Optional[List[tensorflow.python.framework.ops.Tensor]] = None, decoder_hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, decoder_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, cross_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, encoder_last_hidden_state: Optional[tensorflow.python.framework.ops.Tensor] = None, encoder_hidden_states: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, encoder_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None, encoder_global_attentions: Optional[Tuple[tensorflow.python.framework.ops.Tensor]] = None)

Base class for sequence-to-sequence language models outputs.

Parameters

• loss (tf.Tensor of shape (1,), optional, returned when labels is provided) – Language modeling loss.

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

• past_key_values (List[tf.Tensor], optional, returned when use_cache=True is passed or when config.use_cache=True) –

  List of tf.Tensor of length config.n_layers, with each tensor of shape (2, batch_size, num_heads, sequence_length, embed_size_per_head)).

  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• encoder_global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

LEDModel

class transformers.LEDModel(config: transformers.models.led.configuration_led.LEDConfig)

The bare LED Model 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 (LEDConfig) – 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, decoder_input_ids=None, decoder_attention_mask=None, head_mask=None, decoder_head_mask=None, cross_attn_head_mask=None, encoder_outputs=None, global_attention_mask=None, past_key_values=None, inputs_embeds=None, decoder_inputs_embeds=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)

The LEDModel 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. Padding will be ignored by default should you provide it.

  Indices can be obtained using LEDTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

• attention_mask (torch.Tensor of shape (batch_size, sequence_length), optional) –

  Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:

  • 1 for tokens that are not masked,

  • 0 for tokens that are masked.

  What are attention masks?

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

  Indices of decoder input sequence tokens in the vocabulary.

  Indices can be obtained using LedTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

  LED uses the eos_token_id as the starting token for decoder_input_ids generation. If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

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

  Default behavior: generate a tensor that ignores pad tokens in decoder_input_ids. Causal mask will also be used by default.

  If you want to change padding behavior, you should read modeling_led._prepare_decoder_inputs() and modify to your needs. See diagram 1 in the paper for more information on the default strategy.

• global_attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional) –

  Mask to decide the attention given on each token, local attention or global attention for the encoder. Tokens with global attention attends to all other tokens, and all other tokens attend to them. This is important for task-specific finetuning because it makes the model more flexible at representing the task. For example, for classification, the <s> token should be given global attention. For QA, all question tokens should also have global attention. Please refer to the Longformer paper for more details. Mask values selected in [0, 1]:

  • 0 for local attention (a sliding window attention),

  • 1 for global attention (tokens that attend to all other tokens, and all other tokens attend to them).

• head_mask (torch.Tensor of shape (encoder_layers, encoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• decoder_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• cross_attn_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• encoder_outputs (tuple(tuple(torch.FloatTensor), optional) – Tuple consists of (last_hidden_state, optional: hidden_states, optional: attentions) last_hidden_state of shape (batch_size, sequence_length, hidden_size), optional) is a sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder.

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

  Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

  If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_input_ids` of shape (batch_size, sequence_length).

• 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.

• decoder_inputs_embeds (torch.FloatTensor of shape (batch_size, target_sequence_length, hidden_size), optional) –

  Optionally, instead of passing decoder_input_ids you can choose to directly pass an embedded representation. If past_key_values is used, optionally only the last decoder_inputs_embeds have to be input (see past_key_values). This is useful if you want more control over how to convert decoder_input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  If decoder_input_ids and decoder_inputs_embeds are both unset, decoder_inputs_embeds takes the value of inputs_embeds.

• use_cache (bool, optional) – If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).

• 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 Seq2SeqModelOutput 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 (LEDConfig) 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 decoder of the model.

  If past_key_values is used only the last hidden-state of the sequences of shape (batch_size, 1, hidden_size) is output.

• past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) – Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(torch.FloatTensor), optional, returned 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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

Seq2SeqModelOutput or tuple(torch.FloatTensor)

Example:

>>> from transformers import LEDTokenizer, LEDModel
>>> import torch

>>> tokenizer = LEDTokenizer.from_pretrained('allenai/led-base-16384')
>>> model = LEDModel.from_pretrained('allenai/led-base-16384')

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

>>> last_hidden_states = outputs.last_hidden_state

LEDForConditionalGeneration

class transformers.LEDForConditionalGeneration(config: transformers.models.led.configuration_led.LEDConfig)

The LED Model with a language modeling head. Can be used for summarization. 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 (LEDConfig) – 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, decoder_input_ids=None, decoder_attention_mask=None, head_mask=None, decoder_head_mask=None, cross_attn_head_mask=None, encoder_outputs=None, global_attention_mask=None, past_key_values=None, inputs_embeds=None, decoder_inputs_embeds=None, labels=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)

The LEDForConditionalGeneration 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. Padding will be ignored by default should you provide it.

  Indices can be obtained using LEDTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

• attention_mask (torch.Tensor of shape (batch_size, sequence_length), optional) –

  Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:

  • 1 for tokens that are not masked,

  • 0 for tokens that are masked.

  What are attention masks?

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

  Indices of decoder input sequence tokens in the vocabulary.

  Indices can be obtained using LedTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

  LED uses the eos_token_id as the starting token for decoder_input_ids generation. If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

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

  Default behavior: generate a tensor that ignores pad tokens in decoder_input_ids. Causal mask will also be used by default.

  If you want to change padding behavior, you should read modeling_led._prepare_decoder_inputs() and modify to your needs. See diagram 1 in the paper for more information on the default strategy.

• global_attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional) –

  Mask to decide the attention given on each token, local attention or global attention for the encoder. Tokens with global attention attends to all other tokens, and all other tokens attend to them. This is important for task-specific finetuning because it makes the model more flexible at representing the task. For example, for classification, the <s> token should be given global attention. For QA, all question tokens should also have global attention. Please refer to the Longformer paper for more details. Mask values selected in [0, 1]:

  • 0 for local attention (a sliding window attention),

  • 1 for global attention (tokens that attend to all other tokens, and all other tokens attend to them).

• head_mask (torch.Tensor of shape (encoder_layers, encoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• decoder_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• cross_attn_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• encoder_outputs (tuple(tuple(torch.FloatTensor), optional) – Tuple consists of (last_hidden_state, optional: hidden_states, optional: attentions) last_hidden_state of shape (batch_size, sequence_length, hidden_size), optional) is a sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder.

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

  Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

  If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_input_ids` of shape (batch_size, sequence_length).

• 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.

• decoder_inputs_embeds (torch.FloatTensor of shape (batch_size, target_sequence_length, hidden_size), optional) –

  Optionally, instead of passing decoder_input_ids you can choose to directly pass an embedded representation. If past_key_values is used, optionally only the last decoder_inputs_embeds have to be input (see past_key_values). This is useful if you want more control over how to convert decoder_input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  If decoder_input_ids and decoder_inputs_embeds are both unset, decoder_inputs_embeds takes the value of inputs_embeds.

• use_cache (bool, optional) – If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).

• 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 computing the masked language modeling loss. Indices should either be in [0, ..., config.vocab_size] or -100 (see input_ids docstring). Tokens with indices set to -100 are ignored (masked), the loss is only computed for the tokens with labels in [0, ..., config.vocab_size].

Returns

A Seq2SeqLMOutput 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 (LEDConfig) and inputs.

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

• logits (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).

• past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) – Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(torch.FloatTensor), optional, returned 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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

Conditional generation example:

>>> from transformers import LEDTokenizer, LEDForConditionalGeneration
>>> tokenizer = LEDTokenizer.from_pretrained('allenai/led-base-16384')
>>> TXT = "My friends are <mask> but they eat too many carbs."

>>> model = LEDForConditionalGeneration.from_pretrained('allenai/led-base-16384')
>>> input_ids = tokenizer([TXT], return_tensors='pt')['input_ids']

>>> prediction = model.generate(input_ids)[0]
>>> print(tokenizer.decode(prediction, skip_special_tokens=True))

Return type

Seq2SeqLMOutput or tuple(torch.FloatTensor)

Summarization example:

>>> import torch
>>> from transformers import LEDTokenizer, LEDForConditionalGeneration

>>> model = LEDForConditionalGeneration.from_pretrained('allenai/led-large-16384-arxiv')
>>> tokenizer = LEDTokenizer.from_pretrained('allenai/led-large-16384-arxiv')

>>> ARTICLE_TO_SUMMARIZE = '''Transformers (Vaswani et al., 2017) have achieved state-of-the-art
... results in a wide range of natural language tasks including generative
... language modeling (Dai et al., 2019; Radford et al., 2019) and discriminative
... language understanding (Devlin et al., 2019). This success is partly due to
... the self-attention component which enables the network to capture contextual
... information from the entire sequence. While powerful, the memory and computational
... requirements of self-attention grow quadratically with sequence length, making
... it infeasible (or very expensive) to process long sequences.
...
... To address this limitation, we present Longformer, a modified Transformer
... architecture with a self-attention operation that scales linearly with the
... sequence length, making it versatile for processing long documents (Fig 1). This
... is an advantage for natural language tasks such as long document classification,
... question answering (QA), and coreference resolution, where existing approaches
... partition or shorten the long context into smaller sequences that fall within the
... typical 512 token limit of BERT-style pretrained models. Such partitioning could
... potentially result in loss of important cross-partition information, and to
... mitigate this problem, existing methods often rely on complex architectures to
... address such interactions. On the other hand, our proposed Longformer is able to
... build contextual representations of the entire context using multiple layers of
... attention, reducing the need for task-specific architectures.'''
>>> inputs = tokenizer.encode(ARTICLE_TO_SUMMARIZE, return_tensors='pt')

>>> # Global attention on the first token (cf. Beltagy et al. 2020)
>>> global_attention_mask = torch.zeros_like(inputs)
>>> global_attention_mask[:, 0] = 1

>>> # Generate Summary
>>> summary_ids = model.generate(inputs, global_attention_mask=global_attention_mask,
...                              num_beams=3, max_length=32, early_stopping=True)
>>> print(tokenizer.decode(summary_ids[0], skip_special_tokens=True, clean_up_tokenization_spaces=True))

LEDForSequenceClassification

class transformers.LEDForSequenceClassification(config: transformers.models.led.configuration_led.LEDConfig, **kwargs)

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

This model inherits from PreTrainedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)

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

Parameters

config (LEDConfig) – 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, decoder_input_ids=None, decoder_attention_mask=None, head_mask=None, decoder_head_mask=None, cross_attn_head_mask=None, encoder_outputs=None, global_attention_mask=None, inputs_embeds=None, decoder_inputs_embeds=None, labels=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)

The LEDForSequenceClassification 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. Padding will be ignored by default should you provide it.

  Indices can be obtained using LEDTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

• attention_mask (torch.Tensor of shape (batch_size, sequence_length), optional) –

  Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:

  • 1 for tokens that are not masked,

  • 0 for tokens that are masked.

  What are attention masks?

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

  Indices of decoder input sequence tokens in the vocabulary.

  Indices can be obtained using LedTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

  LED uses the eos_token_id as the starting token for decoder_input_ids generation. If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

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

  Default behavior: generate a tensor that ignores pad tokens in decoder_input_ids. Causal mask will also be used by default.

  If you want to change padding behavior, you should read modeling_led._prepare_decoder_inputs() and modify to your needs. See diagram 1 in the paper for more information on the default strategy.

• global_attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional) –

  Mask to decide the attention given on each token, local attention or global attention for the encoder. Tokens with global attention attends to all other tokens, and all other tokens attend to them. This is important for task-specific finetuning because it makes the model more flexible at representing the task. For example, for classification, the <s> token should be given global attention. For QA, all question tokens should also have global attention. Please refer to the Longformer paper for more details. Mask values selected in [0, 1]:

  • 0 for local attention (a sliding window attention),

  • 1 for global attention (tokens that attend to all other tokens, and all other tokens attend to them).

• head_mask (torch.Tensor of shape (encoder_layers, encoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• decoder_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• cross_attn_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• encoder_outputs (tuple(tuple(torch.FloatTensor), optional) – Tuple consists of (last_hidden_state, optional: hidden_states, optional: attentions) last_hidden_state of shape (batch_size, sequence_length, hidden_size), optional) is a sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder.

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

  Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

  If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_input_ids` of shape (batch_size, sequence_length).

• 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.

• decoder_inputs_embeds (torch.FloatTensor of shape (batch_size, target_sequence_length, hidden_size), optional) –

  Optionally, instead of passing decoder_input_ids you can choose to directly pass an embedded representation. If past_key_values is used, optionally only the last decoder_inputs_embeds have to be input (see past_key_values). This is useful if you want more control over how to convert decoder_input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  If decoder_input_ids and decoder_inputs_embeds are both unset, decoder_inputs_embeds takes the value of inputs_embeds.

• use_cache (bool, optional) – If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).

• 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 classification loss is computed (Cross-Entropy).

Returns

A Seq2SeqSequenceClassifierOutput 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 (LEDConfig) and inputs.

• 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).

• past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) – Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(torch.FloatTensor), optional, returned 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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

Seq2SeqSequenceClassifierOutput or tuple(torch.FloatTensor)

Example:

>>> from transformers import LEDTokenizer, LEDForSequenceClassification
>>> import torch

>>> tokenizer = LEDTokenizer.from_pretrained('allenai/led-base-16384')
>>> model = LEDForSequenceClassification.from_pretrained('allenai/led-base-16384')

>>> 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

LEDForQuestionAnswering

class transformers.LEDForQuestionAnswering(config)

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

This model inherits from PreTrainedModel. Check the superclass documentation for the generic methods the library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads etc.)

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

Parameters

config (LEDConfig) – 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, decoder_input_ids=None, decoder_attention_mask=None, head_mask=None, decoder_head_mask=None, cross_attn_head_mask=None, encoder_outputs=None, global_attention_mask=None, start_positions=None, end_positions=None, inputs_embeds=None, decoder_inputs_embeds=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None)

The LEDForQuestionAnswering 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. Padding will be ignored by default should you provide it.

  Indices can be obtained using LEDTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

• attention_mask (torch.Tensor of shape (batch_size, sequence_length), optional) –

  Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:

  • 1 for tokens that are not masked,

  • 0 for tokens that are masked.

  What are attention masks?

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

  Indices of decoder input sequence tokens in the vocabulary.

  Indices can be obtained using LedTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

  LED uses the eos_token_id as the starting token for decoder_input_ids generation. If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

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

  Default behavior: generate a tensor that ignores pad tokens in decoder_input_ids. Causal mask will also be used by default.

  If you want to change padding behavior, you should read modeling_led._prepare_decoder_inputs() and modify to your needs. See diagram 1 in the paper for more information on the default strategy.

• global_attention_mask (torch.FloatTensor of shape (batch_size, sequence_length), optional) –

  Mask to decide the attention given on each token, local attention or global attention for the encoder. Tokens with global attention attends to all other tokens, and all other tokens attend to them. This is important for task-specific finetuning because it makes the model more flexible at representing the task. For example, for classification, the <s> token should be given global attention. For QA, all question tokens should also have global attention. Please refer to the Longformer paper for more details. Mask values selected in [0, 1]:

  • 0 for local attention (a sliding window attention),

  • 1 for global attention (tokens that attend to all other tokens, and all other tokens attend to them).

• head_mask (torch.Tensor of shape (encoder_layers, encoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• decoder_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• cross_attn_head_mask (torch.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• encoder_outputs (tuple(tuple(torch.FloatTensor), optional) – Tuple consists of (last_hidden_state, optional: hidden_states, optional: attentions) last_hidden_state of shape (batch_size, sequence_length, hidden_size), optional) is a sequence of hidden-states at the output of the last layer of the encoder. Used in the cross-attention of the decoder.

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

  Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

  If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_input_ids` of shape (batch_size, sequence_length).

• 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.

• decoder_inputs_embeds (torch.FloatTensor of shape (batch_size, target_sequence_length, hidden_size), optional) –

  Optionally, instead of passing decoder_input_ids you can choose to directly pass an embedded representation. If past_key_values is used, optionally only the last decoder_inputs_embeds have to be input (see past_key_values). This is useful if you want more control over how to convert decoder_input_ids indices into associated vectors than the model’s internal embedding lookup matrix.

  If decoder_input_ids and decoder_inputs_embeds are both unset, decoder_inputs_embeds takes the value of inputs_embeds.

• use_cache (bool, optional) – If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values).

• 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.

• start_positions (torch.LongTensor of shape (batch_size,), optional) – Labels for position (index) of the start of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length). Position outside of the sequence are not taken into account for computing the loss.

• end_positions (torch.LongTensor of shape (batch_size,), optional) – Labels for position (index) of the end of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length). Position outside of the sequence are not taken into account for computing the loss.

Returns

A Seq2SeqQuestionAnsweringModelOutput 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 (LEDConfig) and inputs.

• 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_logits (torch.FloatTensor of shape (batch_size, sequence_length)) – Span-start scores (before SoftMax).

• end_logits (torch.FloatTensor of shape (batch_size, sequence_length)) – Span-end scores (before SoftMax).

• past_key_values (tuple(tuple(torch.FloatTensor)), optional, returned when use_cache=True is passed or when config.use_cache=True) – Tuple of tuple(torch.FloatTensor) of length config.n_layers, with each tuple having 2 tensors of shape (batch_size, num_heads, sequence_length, embed_size_per_head)) and 2 additional tensors of shape (batch_size, num_heads, encoder_sequence_length, embed_size_per_head).

  Contains pre-computed hidden-states (key and values in the self-attention blocks and in the cross-attention blocks) that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(torch.FloatTensor), optional, returned 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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

Return type

Seq2SeqQuestionAnsweringModelOutput or tuple(torch.FloatTensor)

Example:

>>> from transformers import LEDTokenizer, LEDForQuestionAnswering
>>> import torch

>>> tokenizer = LEDTokenizer.from_pretrained('allenai/led-base-16384')
>>> model = LEDForQuestionAnswering.from_pretrained('allenai/led-base-16384')

>>> question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet"
>>> inputs = tokenizer(question, text, return_tensors='pt')
>>> start_positions = torch.tensor([1])
>>> end_positions = torch.tensor([3])

>>> outputs = model(**inputs, start_positions=start_positions, end_positions=end_positions)
>>> loss = outputs.loss
>>> start_scores = outputs.start_logits
>>> end_scores = outputs.end_logits

TFLEDModel

class transformers.TFLEDModel(*args, **kwargs)

The bare LED Model 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(input_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 (LEDConfig) – 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, attention_mask=None, decoder_input_ids=None, decoder_attention_mask=None, head_mask=None, decoder_head_mask=None, encoder_outputs: Optional[Union[Tuple, transformers.models.led.modeling_tf_led.TFLEDEncoderBaseModelOutput]] = None, global_attention_mask=None, past_key_values=None, inputs_embeds=None, decoder_inputs_embeds=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None, training=False, **kwargs)

The TFLEDModel 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 of shape (batch_size, sequence_length)) –

  Indices of input sequence tokens in the vocabulary.

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

  What are input IDs?

• attention_mask (tf.Tensor of shape (batch_size, sequence_length), optional) –

  Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:

  • 1 for tokens that are not masked,

  • 0 for tokens that are masked.

  What are attention masks?

• decoder_input_ids (tf.LongTensor of shape (batch_size, target_sequence_length), optional) –

  Indices of decoder input sequence tokens in the vocabulary.

  Indices can be obtained using LedTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

  LED uses the eos_token_id as the starting token for decoder_input_ids generation. If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

• decoder_attention_mask (tf.Tensor of shape (batch_size, target_sequence_length), optional) – will be made by default and ignore pad tokens. It is not recommended to set this for most use cases.

• head_mask (tf.Tensor of shape (encoder_layers, encoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• decoder_head_mask (tf.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• encoder_outputs (tf.FloatTensor, optional) – hidden states at the output of the last layer of the encoder. Used in the cross-attention of the decoder. of shape (batch_size, sequence_length, hidden_size) is a sequence of

• past_key_values (Tuple[Tuple[tf.Tensor]] of length config.n_layers) – contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding. If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_input_ids of shape (batch_size, sequence_length).

• use_cache (bool, optional, defaults to True) – If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values). Set to False during training, True during generation

• 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 TFLEDSeq2SeqModelOutput 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 (LEDConfig) 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 decoder of the model.

  If past_key_values is used only the last hidden-state of the sequences of shape (batch_size, 1, hidden_size) is output.

• past_key_values (List[tf.Tensor], optional, returned when use_cache=True is passed or when config.use_cache=True) – List of tf.Tensor of length config.n_layers, with each tensor of shape (2, batch_size, num_heads, sequence_length, embed_size_per_head)).

  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• encoder_global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

Return type

TFLEDSeq2SeqModelOutput or tuple(tf.Tensor)

Example:

>>> from transformers import LEDTokenizer, TFLEDModel
>>> import tensorflow as tf

>>> tokenizer = LEDTokenizer.from_pretrained('allenai/led-base-16384')
>>> model = TFLEDModel.from_pretrained('allenai/led-base-16384')

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

>>> last_hidden_states = outputs.last_hidden_state

TFLEDForConditionalGeneration

class transformers.TFLEDForConditionalGeneration(*args, **kwargs)

The LED Model with a language modeling head. Can be used for summarization. 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(input_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 (LEDConfig) – 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, attention_mask=None, decoder_input_ids=None, decoder_attention_mask=None, head_mask=None, decoder_head_mask=None, encoder_outputs: Optional[transformers.models.led.modeling_tf_led.TFLEDEncoderBaseModelOutput] = None, global_attention_mask=None, past_key_values=None, inputs_embeds=None, decoder_inputs_embeds=None, use_cache=None, output_attentions=None, output_hidden_states=None, return_dict=None, labels=None, training=False, **kwargs)

The TFLEDForConditionalGeneration 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 of shape ({0})) –

  Indices of input sequence tokens in the vocabulary.

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

  What are input IDs?

• attention_mask (tf.Tensor of shape ({0}), optional) –

  Mask to avoid performing attention on padding token indices. Mask values selected in [0, 1]:

  • 1 for tokens that are not masked,

  • 0 for tokens that are masked.

  What are attention masks?

• decoder_input_ids (tf.LongTensor of shape (batch_size, target_sequence_length), optional) –

  Indices of decoder input sequence tokens in the vocabulary.

  Indices can be obtained using LedTokenizer. See transformers.PreTrainedTokenizer.encode() and transformers.PreTrainedTokenizer.__call__() for details.

  What are input IDs?

  LED uses the eos_token_id as the starting token for decoder_input_ids generation. If past_key_values is used, optionally only the last decoder_input_ids have to be input (see past_key_values).

• decoder_attention_mask (tf.Tensor of shape (batch_size, target_sequence_length), optional) – will be made by default and ignore pad tokens. It is not recommended to set this for most use cases.

• head_mask (tf.Tensor of shape (encoder_layers, encoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• decoder_head_mask (tf.Tensor of shape (decoder_layers, decoder_attention_heads), optional) –

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

  • 1 indicates the head is not masked,

  • 0 indicates the head is masked.

• encoder_outputs (tf.FloatTensor, optional) – hidden states at the output of the last layer of the encoder. Used in the cross-attention of the decoder. of shape (batch_size, sequence_length, hidden_size) is a sequence of

• past_key_values (Tuple[Tuple[tf.Tensor]] of length config.n_layers) – contains precomputed key and value hidden states of the attention blocks. Can be used to speed up decoding. If past_key_values are used, the user can optionally input only the last decoder_input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all decoder_input_ids of shape (batch_size, sequence_length).

• use_cache (bool, optional, defaults to True) – If set to True, past_key_values key value states are returned and can be used to speed up decoding (see past_key_values). Set to False during training, True during generation

• 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 TFLEDSeq2SeqLMOutput 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 (LEDConfig) and inputs.

• loss (tf.Tensor of shape (1,), optional, returned when labels is provided) – Language modeling loss.

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

• past_key_values (List[tf.Tensor], optional, returned when use_cache=True is passed or when config.use_cache=True) – List of tf.Tensor of length config.n_layers, with each tensor of shape (2, batch_size, num_heads, sequence_length, embed_size_per_head)).

  Contains pre-computed hidden-states (key and values in the attention blocks) of the decoder that can be used (see past_key_values input) to speed up sequential decoding.

• decoder_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 decoder at the output of each layer plus the initial embedding outputs.

• decoder_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 of the decoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• cross_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 of the decoder’s cross-attention layer, after the attention softmax, used to compute the weighted average in the cross-attention heads.

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

• encoder_hidden_states (tuple(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 encoder at the output of each layer plus the initial embedding outputs.

• encoder_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 of the encoder, after the attention softmax, used to compute the weighted average in the self-attention heads.

• encoder_global_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, x), where x is the number of tokens with global attention mask.

  Global attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads. Those are the attention weights from every token with global attention to every token in the sequence.

Examples:

>>> from transformers import LEDTokenizer, TFLEDForConditionalGeneration
>>> import tensorflow as tf
>>> mname = 'allenai/led-base-16384'
>>> tokenizer = LEDTokenizer.from_pretrained(mname)
>>> TXT = "My friends are <mask> but they eat too many carbs."
>>> model = TFLEDForConditionalGeneration.from_pretrained(mname)
>>> batch = tokenizer([TXT], return_tensors='tf')
>>> logits = model(inputs=batch.input_ids).logits
>>> probs = tf.nn.softmax(logits[0])
>>> # probs[5] is associated with the mask token

Return type

TFLEDSeq2SeqLMOutput or tuple(tf.Tensor)