CANINE¶
Overview¶
The CANINE model was proposed in CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting. It’s among the first papers that trains a Transformer without using an explicit tokenization step (such as Byte Pair Encoding (BPE), WordPiece or SentencePiece). Instead, the model is trained directly at a Unicode character-level. Training at a character-level inevitably comes with a longer sequence length, which CANINE solves with an efficient downsampling strategy, before applying a deep Transformer encoder.
The abstract from the paper is the following:
Pipelined NLP systems have largely been superseded by end-to-end neural modeling, yet nearly all commonly-used models still require an explicit tokenization step. While recent tokenization approaches based on data-derived subword lexicons are less brittle than manually engineered tokenizers, these techniques are not equally suited to all languages, and the use of any fixed vocabulary may limit a model’s ability to adapt. In this paper, we present CANINE, a neural encoder that operates directly on character sequences, without explicit tokenization or vocabulary, and a pre-training strategy that operates either directly on characters or optionally uses subwords as a soft inductive bias. To use its finer-grained input effectively and efficiently, CANINE combines downsampling, which reduces the input sequence length, with a deep transformer stack, which encodes context. CANINE outperforms a comparable mBERT model by 2.8 F1 on TyDi QA, a challenging multilingual benchmark, despite having 28% fewer model parameters.
Tips:
CANINE uses no less than 3 Transformer encoders internally: 2 “shallow” encoders (which only consist of a single layer) and 1 “deep” encoder (which is a regular BERT encoder). First, a “shallow” encoder is used to contextualize the character embeddings, using local attention. Next, after downsampling, a “deep” encoder is applied. Finally, after upsampling, a “shallow” encoder is used to create the final character embeddings. Details regarding up- and downsampling can be found in the paper.
CANINE uses a max sequence length of 2048 characters by default. One can use
CanineTokenizer
to prepare text for the model.Classification can be done by placing a linear layer on top of the final hidden state of the special [CLS] token (which has a predefined Unicode code point). For token classification tasks however, the downsampled sequence of tokens needs to be upsampled again to match the length of the original character sequence (which is 2048). The details for this can be found in the paper.
Models:
google/canine-c: Pre-trained with autoregressive character loss, 12-layer, 768-hidden, 12-heads, 121M parameters (size ~500 MB).
google/canine-s: Pre-trained with subword loss, 12-layer, 768-hidden, 12-heads, 121M parameters (size ~500 MB).
This model was contributed by nielsr. The original code can be found here.
Example¶
CANINE works on raw characters, so it can be used without a tokenizer:
from transformers import CanineModel
import torch
model = CanineModel.from_pretrained('google/canine-c') # model pre-trained with autoregressive character loss
text = "hello world"
# use Python's built-in ord() function to turn each character into its unicode code point id
input_ids = torch.tensor([[ord(char) for char in text]])
outputs = model(input_ids) # forward pass
pooled_output = outputs.pooler_output
sequence_output = outputs.last_hidden_state
For batched inference and training, it is however recommended to make use of the tokenizer (to pad/truncate all sequences to the same length):
from transformers import CanineTokenizer, CanineModel
model = CanineModel.from_pretrained('google/canine-c')
tokenizer = CanineTokenizer.from_pretrained('google/canine-c')
inputs = ["Life is like a box of chocolates.", "You never know what you gonna get."]
encoding = tokenizer(inputs, padding="longest", truncation=True, return_tensors="pt")
outputs = model(**encoding) # forward pass
pooled_output = outputs.pooler_output
sequence_output = outputs.last_hidden_state
CANINE specific outputs¶
-
class
transformers.models.canine.modeling_canine.
CanineModelOutputWithPooling
(last_hidden_state: Optional[torch.FloatTensor] = None, pooler_output: Optional[torch.FloatTensor] = None, hidden_states: Optional[Tuple[torch.FloatTensor]] = None, attentions: Optional[Tuple[torch.FloatTensor]] = None)[source]¶ Output type of
CanineModel
. Based onBaseModelOutputWithPooling
, but with slightly differenthidden_states
andattentions
, as these also include the hidden states and attentions of the shallow Transformer encoders.- 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 (i.e. the output of the final shallow Transformer encoder).pooler_output (
torch.FloatTensor
of shape(batch_size, hidden_size)
) – Hidden-state of the first token of the sequence (classification token) at the last layer of the deep Transformer encoder, further processed by a Linear layer and a Tanh activation function. The Linear layer weights are trained from the next sentence prediction (classification) objective during pretraining.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) – Tuple oftorch.FloatTensor
(one for the input to each encoder + one for the output of each layer of each encoder) of shape(batch_size, sequence_length, hidden_size)
and(batch_size, sequence_length // config.downsampling_rate, hidden_size)
. Hidden-states of the model at the output of each layer plus the initial input to each Transformer encoder. The hidden states of the shallow encoders have lengthsequence_length
, but the hidden states of the deep encoder have lengthsequence_length
//config.downsampling_rate
.attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) – Tuple oftorch.FloatTensor
(one for each layer) of the 3 Transformer encoders of shape(batch_size, num_heads, sequence_length, sequence_length)
and(batch_size, num_heads, sequence_length // config.downsampling_rate, sequence_length // config.downsampling_rate)
. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
CanineConfig¶
-
class
transformers.
CanineConfig
(hidden_size=768, num_hidden_layers=12, num_attention_heads=12, intermediate_size=3072, hidden_act='gelu', hidden_dropout_prob=0.1, attention_probs_dropout_prob=0.1, max_position_embeddings=16384, type_vocab_size=16, initializer_range=0.02, layer_norm_eps=1e-12, use_cache=True, is_encoder_decoder=False, pad_token_id=0, bos_token_id=57344, eos_token_id=57345, downsampling_rate=4, upsampling_kernel_size=4, num_hash_functions=8, num_hash_buckets=16384, local_transformer_stride=128, **kwargs)[source]¶ This is the configuration class to store the configuration of a
CanineModel
. It is used to instantiate an CANINE 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 CANINE google/canine-s architecture.Configuration objects inherit from
PretrainedConfig
and can be used to control the model outputs. Read the documentation fromPretrainedConfig
for more information.- Parameters
hidden_size (
int
, optional, defaults to 768) – Dimension of the encoder layers and the pooler layer.num_hidden_layers (
int
, optional, defaults to 12) – Number of hidden layers in the deep Transformer encoder.num_attention_heads (
int
, optional, defaults to 12) – Number of attention heads for each attention layer in the Transformer encoders.intermediate_size (
int
, optional, defaults to 3072) – Dimension of the “intermediate” (i.e., feed-forward) layer in the Transformer encoders.hidden_act (
str
orfunction
, optional, defaults to"gelu"
) – The non-linear activation function (function or string) in the encoder and pooler. If string,"gelu"
,"relu"
,"selu"
and"gelu_new"
are supported.hidden_dropout_prob (
float
, optional, defaults to 0.1) – The dropout probabilitiy for all fully connected layers in the embeddings, encoders, and pooler.attention_probs_dropout_prob (
float
, optional, defaults to 0.1) – The dropout ratio for the attention probabilities.max_position_embeddings (
int
, optional, defaults to 16384) – The maximum sequence length that this model might ever be used with.type_vocab_size (
int
, optional, defaults to 16) – The vocabulary size of thetoken_type_ids
passed when callingCanineModel
.initializer_range (
float
, optional, defaults to 0.02) – The standard deviation of the truncated_normal_initializer for initializing all weight matrices.layer_norm_eps (
float
, optional, defaults to 1e-12) – The epsilon used by the layer normalization layers.gradient_checkpointing (
bool
, optional, defaults toFalse
) – IfTrue
, use gradient checkpointing to save memory at the expense of slower backward pass.downsampling_rate (
int
, optional, defaults to 4) – The rate at which to downsample the original character sequence length before applying the deep Transformer encoder.upsampling_kernel_size (
int
, optional, defaults to 4) – The kernel size (i.e. the number of characters in each window) of the convolutional projection layer when projecting back fromhidden_size`*2 to :obj:`hidden_size
.num_hash_functions (
int
, optional, defaults to 8) – The number of hash functions to use. Each hash function has its own embedding matrix.num_hash_buckets (
int
, optional, defaults to 16384) – The number of hash buckets to use.local_transformer_stride (
int
, optional, defaults to 128) – The stride of the local attention of the first shallow Transformer encoder. Defaults to 128 for good TPU/XLA memory alignment.
Example:
>>> from transformers import CanineModel, CanineConfig >>> # Initializing a CANINE google/canine-s style configuration >>> configuration = CanineConfig() >>> # Initializing a model from the google/canine-s style configuration >>> model = CanineModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config
CanineTokenizer¶
-
class
transformers.
CanineTokenizer
(bos_token='\ue000', eos_token='\ue001', sep_token='\ue001', cls_token='\ue000', pad_token='\x00', mask_token='\ue003', add_prefix_space=False, model_max_length=2048, **kwargs)[source]¶ Construct a CANINE tokenizer (i.e. a character splitter). It turns text into a sequence of characters, and then converts each character into its Unicode code point.
CanineTokenizer
inherits fromPreTrainedTokenizer
.Refer to superclass
PreTrainedTokenizer
for usage examples and documentation concerning parameters.- Parameters
model_max_length (
int
, optional, defaults to 2048) – The maximum sentence length the model accepts.
-
build_inputs_with_special_tokens
(token_ids_0: List[int], token_ids_1: Optional[List[int]] = None) → List[int][source]¶ Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens. A CANINE sequence has the following format:
single sequence:
[CLS] X [SEP]
pair of sequences:
[CLS] A [SEP] B [SEP]
- 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][source]¶ Create a mask from the two sequences passed to be used in a sequence-pair classification task. A CANINE sequence pair mask has the following format:
0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 | first sequence | second sequence |
If
token_ids_1
isNone
, this method only returns the first portion of the mask (0s).- Parameters
token_ids_0 (
List[int]
) – List of IDs.token_ids_1 (
List[int]
, optional) – Optional second list of IDs for sequence pairs.
- Returns
List of token type IDs according to the given sequence(s).
- Return type
List[int]
-
get_special_tokens_mask
(token_ids_0: List[int], token_ids_1: Optional[List[int]] = None, already_has_special_tokens: bool = False) → List[int][source]¶ 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 toFalse
) – 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]
CanineModel¶
-
class
transformers.
CanineModel
(config, add_pooling_layer=True)[source]¶ The bare CANINE Model transformer outputting raw hidden-states without any specific head on top. This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
CanineConfig
) – Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out thefrom_pretrained()
method to load the model weights.
-
forward
(input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]¶ The
CanineModel
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) –Indices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.CanineTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
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.
token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) –Segment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0 corresponds to a sentence A token,
1 corresponds to a sentence B token.
position_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) –Indices of positions of each input sequence tokens in the position embeddings. Selected in the range
[0, config.max_position_embeddings - 1]
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) –Mask to nullify selected heads of the self-attention modules. Mask values selected in
[0, 1]
:1 indicates the head is not masked,
0 indicates the head is masked.
inputs_embeds (
torch.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) – Optionally, instead of passinginput_ids
you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.output_attentions (
bool
, optional) – Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) – Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) – Whether or not to return aModelOutput
instead of a plain tuple.
- Returns
A
CanineModelOutputWithPooling
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) comprising various elements depending on the configuration (CanineConfig
) and inputs.last_hidden_state (
torch.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
) – Sequence of hidden-states at the output of the last layer of the model (i.e. the output of the final shallow Transformer encoder).pooler_output (
torch.FloatTensor
of shape(batch_size, hidden_size)
) – Hidden-state of the first token of the sequence (classification token) at the last layer of the deep Transformer encoder, further processed by a Linear layer and a Tanh activation function. The Linear layer weights are trained from the next sentence prediction (classification) objective during pretraining.hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) – Tuple oftorch.FloatTensor
(one for the input to each encoder + one for the output of each layer of each encoder) of shape(batch_size, sequence_length, hidden_size)
and(batch_size, sequence_length // config.downsampling_rate, hidden_size)
. Hidden-states of the model at the output of each layer plus the initial input to each Transformer encoder. The hidden states of the shallow encoders have lengthsequence_length
, but the hidden states of the deep encoder have lengthsequence_length
//config.downsampling_rate
.attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) – Tuple oftorch.FloatTensor
(one for each layer) of the 3 Transformer encoders of shape(batch_size, num_heads, sequence_length, sequence_length)
and(batch_size, num_heads, sequence_length // config.downsampling_rate, sequence_length // config.downsampling_rate)
. Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
CanineModelOutputWithPooling
ortuple(torch.FloatTensor)
Example:
>>> from transformers import CanineTokenizer, CanineModel >>> import torch >>> tokenizer = CanineTokenizer.from_pretrained('google/canine-s') >>> model = CanineModel.from_pretrained('google/canine-s') >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> outputs = model(**inputs) >>> last_hidden_states = outputs.last_hidden_state
CanineForSequenceClassification¶
-
class
transformers.
CanineForSequenceClassification
(config)[source]¶ CANINE Model transformer with a sequence classification/regression head on top (a linear layer on top of the pooled output) e.g. for GLUE tasks.
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
CanineConfig
) – Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out thefrom_pretrained()
method to load the model weights.
-
forward
(input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, labels=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]¶ The
CanineForSequenceClassification
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shapebatch_size, sequence_length
) –Indices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.CanineTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shapebatch_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.
token_type_ids (
torch.LongTensor
of shapebatch_size, sequence_length
, optional) –Segment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0 corresponds to a sentence A token,
1 corresponds to a sentence B token.
position_ids (
torch.LongTensor
of shapebatch_size, sequence_length
, optional) –Indices of positions of each input sequence tokens in the position embeddings. Selected in the range
[0, config.max_position_embeddings - 1]
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) –Mask to nullify selected heads of the self-attention modules. Mask values selected in
[0, 1]
:1 indicates the head is not masked,
0 indicates the head is masked.
inputs_embeds (
torch.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) – Optionally, instead of passinginput_ids
you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.output_attentions (
bool
, optional) – Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) – Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) – Whether or not to return aModelOutput
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]
. Ifconfig.num_labels == 1
a regression loss is computed (Mean-Square loss), Ifconfig.num_labels > 1
a classification loss is computed (Cross-Entropy).
- Returns
A
SequenceClassifierOutput
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) comprising various elements depending on the configuration (CanineConfig
) and inputs.loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabels
is provided) – Classification (or regression if config.num_labels==1) loss.logits (
torch.FloatTensor
of shape(batch_size, config.num_labels)
) – Classification (or regression if config.num_labels==1) scores (before SoftMax).hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) – Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) – Tuple oftorch.FloatTensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
SequenceClassifierOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import CanineTokenizer, CanineForSequenceClassification >>> import torch >>> tokenizer = CanineTokenizer.from_pretrained('google/canine-s') >>> model = CanineForSequenceClassification.from_pretrained('google/canine-s') >>> 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
CanineForMultipleChoice¶
-
class
transformers.
CanineForMultipleChoice
(config)[source]¶ CANINE Model with a multiple choice classification head on top (a linear layer on top of the pooled output and a softmax) e.g. for RocStories/SWAG tasks.
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
CanineConfig
) – Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out thefrom_pretrained()
method to load the model weights.
-
forward
(input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, labels=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]¶ The
CanineForMultipleChoice
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shapebatch_size, num_choices, sequence_length
) –Indices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.CanineTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
of shapebatch_size, num_choices, 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.
token_type_ids (
torch.LongTensor
of shapebatch_size, num_choices, sequence_length
, optional) –Segment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0 corresponds to a sentence A token,
1 corresponds to a sentence B token.
position_ids (
torch.LongTensor
of shapebatch_size, num_choices, sequence_length
, optional) –Indices of positions of each input sequence tokens in the position embeddings. Selected in the range
[0, config.max_position_embeddings - 1]
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) –Mask to nullify selected heads of the self-attention modules. Mask values selected in
[0, 1]
:1 indicates the head is not masked,
0 indicates the head is masked.
inputs_embeds (
torch.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) – Optionally, instead of passinginput_ids
you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.output_attentions (
bool
, optional) – Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) – Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) – Whether or not to return aModelOutput
instead of a plain tuple.labels (
torch.LongTensor
of shape(batch_size,)
, optional) – Labels for computing the multiple choice classification loss. Indices should be in[0, ..., num_choices-1]
wherenum_choices
is the size of the second dimension of the input tensors. (Seeinput_ids
above)
- Returns
A
MultipleChoiceModelOutput
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) comprising various elements depending on the configuration (CanineConfig
) and inputs.loss (
torch.FloatTensor
of shape (1,), optional, returned whenlabels
is provided) – Classification loss.logits (
torch.FloatTensor
of shape(batch_size, num_choices)
) – num_choices is the second dimension of the input tensors. (see input_ids above).Classification scores (before SoftMax).
hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) – Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) – Tuple oftorch.FloatTensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
MultipleChoiceModelOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import CanineTokenizer, CanineForMultipleChoice >>> import torch >>> tokenizer = CanineTokenizer.from_pretrained('google/canine-s') >>> model = CanineForMultipleChoice.from_pretrained('google/canine-s') >>> prompt = "In Italy, pizza served in formal settings, such as at a restaurant, is presented unsliced." >>> choice0 = "It is eaten with a fork and a knife." >>> choice1 = "It is eaten while held in the hand." >>> labels = torch.tensor(0).unsqueeze(0) # choice0 is correct (according to Wikipedia ;)), batch size 1 >>> encoding = tokenizer([prompt, prompt], [choice0, choice1], return_tensors='pt', padding=True) >>> outputs = model(**{k: v.unsqueeze(0) for k,v in encoding.items()}, labels=labels) # batch size is 1 >>> # the linear classifier still needs to be trained >>> loss = outputs.loss >>> logits = outputs.logits
CanineForTokenClassification¶
-
class
transformers.
CanineForTokenClassification
(config)[source]¶ CANINE Model with a token classification head on top (a linear layer on top of the hidden-states output) e.g. for Named-Entity-Recognition (NER) tasks.
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
CanineConfig
) – Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out thefrom_pretrained()
method to load the model weights.
-
forward
(input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, labels=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]¶ The
CanineForTokenClassification
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) –Indices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.CanineTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
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.
token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) –Segment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0 corresponds to a sentence A token,
1 corresponds to a sentence B token.
position_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) –Indices of positions of each input sequence tokens in the position embeddings. Selected in the range
[0, config.max_position_embeddings - 1]
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) –Mask to nullify selected heads of the self-attention modules. Mask values selected in
[0, 1]
:1 indicates the head is not masked,
0 indicates the head is masked.
inputs_embeds (
torch.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) – Optionally, instead of passinginput_ids
you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.output_attentions (
bool
, optional) – Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) – Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) – Whether or not to return aModelOutput
instead of a plain tuple.labels (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) – Labels for computing the token classification loss. Indices should be in[0, ..., config.num_labels - 1]
.
- Returns
A
TokenClassifierOutput
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) comprising various elements depending on the configuration (CanineConfig
) and inputs.loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabels
is provided) – Classification loss.logits (
torch.FloatTensor
of shape(batch_size, sequence_length, config.num_labels)
) – Classification scores (before SoftMax).hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) – Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) – Tuple oftorch.FloatTensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
TokenClassifierOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import CanineTokenizer, CanineForTokenClassification >>> import torch >>> tokenizer = CanineTokenizer.from_pretrained('google/canine-s') >>> model = CanineForTokenClassification.from_pretrained('google/canine-s') >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> labels = torch.tensor([1] * inputs["input_ids"].size(1)).unsqueeze(0) # Batch size 1 >>> outputs = model(**inputs, labels=labels) >>> loss = outputs.loss >>> logits = outputs.logits
CanineForQuestionAnswering¶
-
class
transformers.
CanineForQuestionAnswering
(config)[source]¶ CANINE Model with a span classification head on top for extractive question-answering tasks like SQuAD (a linear layers on top of the hidden-states output to compute span start logits and span end logits).
This model is a PyTorch torch.nn.Module sub-class. Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general usage and behavior.
- Parameters
config (
CanineConfig
) – Model configuration class with all the parameters of the model. Initializing with a config file does not load the weights associated with the model, only the configuration. Check out thefrom_pretrained()
method to load the model weights.
-
forward
(input_ids=None, attention_mask=None, token_type_ids=None, position_ids=None, head_mask=None, inputs_embeds=None, start_positions=None, end_positions=None, output_attentions=None, output_hidden_states=None, return_dict=None)[source]¶ The
CanineForQuestionAnswering
forward method, overrides the__call__()
special method.Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them.- Parameters
input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) –Indices of input sequence tokens in the vocabulary.
Indices can be obtained using
transformers.CanineTokenizer
. Seetransformers.PreTrainedTokenizer.encode()
andtransformers.PreTrainedTokenizer.__call__()
for details.attention_mask (
torch.FloatTensor
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.
token_type_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) –Segment token indices to indicate first and second portions of the inputs. Indices are selected in
[0, 1]
:0 corresponds to a sentence A token,
1 corresponds to a sentence B token.
position_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
, optional) –Indices of positions of each input sequence tokens in the position embeddings. Selected in the range
[0, config.max_position_embeddings - 1]
.head_mask (
torch.FloatTensor
of shape(num_heads,)
or(num_layers, num_heads)
, optional) –Mask to nullify selected heads of the self-attention modules. Mask values selected in
[0, 1]
:1 indicates the head is not masked,
0 indicates the head is masked.
inputs_embeds (
torch.FloatTensor
of shape(batch_size, sequence_length, hidden_size)
, optional) – Optionally, instead of passinginput_ids
you can choose to directly pass an embedded representation. This is useful if you want more control over how to convert input_ids indices into associated vectors than the model’s internal embedding lookup matrix.output_attentions (
bool
, optional) – Whether or not to return the attentions tensors of all attention layers. Seeattentions
under returned tensors for more detail.output_hidden_states (
bool
, optional) – Whether or not to return the hidden states of all layers. Seehidden_states
under returned tensors for more detail.return_dict (
bool
, optional) – Whether or not to return aModelOutput
instead of a plain tuple.start_positions (
torch.LongTensor
of shape(batch_size,)
, optional) – Labels for position (index) of the start of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length
). Position outside of the sequence are not taken into account for computing the loss.end_positions (
torch.LongTensor
of shape(batch_size,)
, optional) – Labels for position (index) of the end of the labelled span for computing the token classification loss. Positions are clamped to the length of the sequence (sequence_length
). Position outside of the sequence are not taken into account for computing the loss.
- Returns
A
QuestionAnsweringModelOutput
or a tuple oftorch.FloatTensor
(ifreturn_dict=False
is passed or whenconfig.return_dict=False
) comprising various elements depending on the configuration (CanineConfig
) and inputs.loss (
torch.FloatTensor
of shape(1,)
, optional, returned whenlabels
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).hidden_states (
tuple(torch.FloatTensor)
, optional, returned whenoutput_hidden_states=True
is passed or whenconfig.output_hidden_states=True
) – Tuple oftorch.FloatTensor
(one for the output of the embeddings + one for the output of each layer) of shape(batch_size, sequence_length, hidden_size)
.Hidden-states of the model at the output of each layer plus the initial embedding outputs.
attentions (
tuple(torch.FloatTensor)
, optional, returned whenoutput_attentions=True
is passed or whenconfig.output_attentions=True
) – Tuple oftorch.FloatTensor
(one for each layer) of shape(batch_size, num_heads, sequence_length, sequence_length)
.Attentions weights after the attention softmax, used to compute the weighted average in the self-attention heads.
- Return type
QuestionAnsweringModelOutput
ortuple(torch.FloatTensor)
Example:
>>> from transformers import CanineTokenizer, CanineForQuestionAnswering >>> import torch >>> tokenizer = CanineTokenizer.from_pretrained('google/canine-s') >>> model = CanineForQuestionAnswering.from_pretrained('google/canine-s') >>> 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