CANINE
Overview
CANINE モデルは、CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation、Jonathan H. Clark、Dan Garrette、Iulia Turc、John Wieting 著。その 明示的なトークン化ステップ (バイト ペアなど) を使用せずに Transformer をトレーニングする最初の論文の 1 つ エンコーディング (BPE、WordPiece または SentencePiece)。代わりに、モデルは Unicode 文字レベルで直接トレーニングされます。 キャラクターレベルでのトレーニングでは必然的にシーケンスの長さが長くなりますが、CANINE はこれを効率的な方法で解決します。 ディープ Transformer エンコーダを適用する前に、ダウンサンプリング戦略を実行します。
論文の要約は次のとおりです。
パイプライン NLP システムは、エンドツーエンドのニューラル モデリングに大部分が取って代わられていますが、一般的に使用されているほぼすべてのモデルは 依然として明示的なトークン化手順が必要です。最近のトークン化アプローチはデータ由来のサブワードに基づいていますが、 レキシコンは手動で作成されたトークナイザーよりも脆弱ではありませんが、これらの技術はすべての言語に等しく適しているわけではありません。 言語や固定語彙の使用により、モデルの適応能力が制限される可能性があります。この論文では、CANINE を紹介します。 明示的なトークン化や語彙を使用せずに、文字シーケンスを直接操作するニューラル エンコーダーと、 文字に直接作用するか、オプションでサブワードをソフト誘導バイアスとして使用する事前トレーニング戦略。 よりきめの細かい入力を効果的かつ効率的に使用するために、CANINE はダウンサンプリングを組み合わせて、入力を削減します。 コンテキストをエンコードするディープトランスフォーマースタックを備えたシーケンスの長さ。 CANINE は、同等の mBERT モデルよりも次の点で優れています。 TyDi QA の 2.8 F1 は、モデル パラメータが 28% 少ないにもかかわらず、困難な多言語ベンチマークです。
このモデルは、nielsr によって提供されました。元のコードは ここ にあります。
Usage tips
- CANINE は内部で少なくとも 3 つの Transformer エンコーダーを使用します: 2 つの「浅い」エンコーダー (単一のエンコーダーのみで構成) レイヤー) と 1 つの「ディープ」エンコーダー (通常の BERT エンコーダー)。まず、「浅い」エンコーダを使用してコンテキストを設定します。 ローカル アテンションを使用した文字の埋め込み。次に、ダウンサンプリングの後、「ディープ」エンコーダーが適用されます。ついに、 アップサンプリング後、「浅い」エンコーダを使用して最終的な文字埋め込みが作成されます。アップと ダウンサンプリングについては論文に記載されています。
- CANINE は、デフォルトで 2048 文字の最大シーケンス長を使用します。 CanineTokenizer を使用できます モデル用のテキストを準備します。
- 特別な [CLS] トークンの最終的な非表示状態の上に線形レイヤーを配置することで分類を行うことができます。 (事前定義された Unicode コード ポイントがあります)。ただし、トークン分類タスクの場合は、ダウンサンプリングされたシーケンス トークンは、元の文字シーケンスの長さ (2048) と一致するように再度アップサンプリングする必要があります。の 詳細については、論文を参照してください。
モデルのチェックポイント:
- google/canine-c: 自己回帰文字損失で事前トレーニング済み、 12 レイヤー、768 隠し、12 ヘッド、121M パラメーター (サイズ ~500 MB)。
- google/canine-s: サブワード損失で事前トレーニング済み、12 層、 768 個の非表示、12 ヘッド、121M パラメーター (サイズ ~500 MB)。
Usage example
CANINE は生の文字で動作するため、トークナイザーなしで使用できます。
>>> 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
ただし、バッチ推論とトレーニングの場合は、トークナイザーを使用することをお勧めします(すべてをパディング/切り詰めるため) シーケンスを同じ長さにします):
>>> 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
Resources
CanineConfig
class transformers.CanineConfig
< source >( 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 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 )
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 calling CanineModel. - 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. - pad_token_id (
int
, optional, defaults to 0) — Padding token id. - bos_token_id (
int
, optional, defaults to 57344) — Beginning of stream token id. - eos_token_id (
int
, optional, defaults to 57345) — End of stream token id. - 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 tohidden_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.
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 from PretrainedConfig for more information.
Example:
>>> from transformers import CanineConfig, CanineModel
>>> # Initializing a CANINE google/canine-s style configuration
>>> configuration = CanineConfig()
>>> # Initializing a model (with random weights) from the google/canine-s style configuration
>>> model = CanineModel(configuration)
>>> # Accessing the model configuration
>>> configuration = model.config
CanineTokenizer
class transformers.CanineTokenizer
< source >( 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 )
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 from PreTrainedTokenizer.
Refer to superclass PreTrainedTokenizer for usage examples and documentation concerning parameters.
build_inputs_with_special_tokens
< source >( token_ids_0: typing.List[int] token_ids_1: typing.Optional[typing.List[int]] = None ) → List[int]
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[int]
List of input IDs with the appropriate special tokens.
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]
get_special_tokens_mask
< source >( token_ids_0: typing.List[int] token_ids_1: typing.Optional[typing.List[int]] = None already_has_special_tokens: bool = False ) → List[int]
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
List[int]
A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
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.
create_token_type_ids_from_sequences
< source >( token_ids_0: typing.List[int] token_ids_1: typing.Optional[typing.List[int]] = None ) → List[int]
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[int]
List of token type IDs according to the given sequence(s).
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
is None
, this method only returns the first portion of the mask (0s).
CANINE specific outputs
class transformers.models.canine.modeling_canine.CanineModelOutputWithPooling
< source >( last_hidden_state: FloatTensor = None pooler_output: FloatTensor = None hidden_states: typing.Optional[typing.Tuple[torch.FloatTensor]] = None attentions: typing.Optional[typing.Tuple[torch.FloatTensor]] = None )
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.
Output type of CanineModel. Based on BaseModelOutputWithPooling, but with slightly
different hidden_states
and attentions
, as these also include the hidden states and attentions of the shallow
Transformer encoders.
CanineModel
class transformers.CanineModel
< source >( config add_pooling_layer = True )
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 the from_pretrained() method to load the model weights.
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.
forward
< source >( input_ids: typing.Optional[torch.LongTensor] = None attention_mask: typing.Optional[torch.FloatTensor] = None token_type_ids: typing.Optional[torch.LongTensor] = None position_ids: typing.Optional[torch.LongTensor] = None head_mask: typing.Optional[torch.FloatTensor] = None inputs_embeds: typing.Optional[torch.FloatTensor] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.models.canine.modeling_canine.CanineModelOutputWithPooling or tuple(torch.FloatTensor)
Parameters
- input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) — Indices of input sequence tokens in the vocabulary.Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and 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 a ModelOutput instead of a plain tuple.
Returns
transformers.models.canine.modeling_canine.CanineModelOutputWithPooling or tuple(torch.FloatTensor)
A transformers.models.canine.modeling_canine.CanineModelOutputWithPooling 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 (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.
The CanineModel forward method, overrides the __call__
special method.
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.
Example:
>>> from transformers import AutoTokenizer, CanineModel
>>> import torch
>>> tokenizer = AutoTokenizer.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
< source >( config )
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 the from_pretrained() method to load the model weights.
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.
forward
< source >( input_ids: typing.Optional[torch.LongTensor] = None attention_mask: typing.Optional[torch.FloatTensor] = None token_type_ids: typing.Optional[torch.LongTensor] = None position_ids: typing.Optional[torch.LongTensor] = None head_mask: typing.Optional[torch.FloatTensor] = None inputs_embeds: typing.Optional[torch.FloatTensor] = None labels: typing.Optional[torch.LongTensor] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.modeling_outputs.SequenceClassifierOutput or tuple(torch.FloatTensor)
Parameters
- input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) — Indices of input sequence tokens in the vocabulary.Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and 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 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]
. Ifconfig.num_labels == 1
a regression loss is computed (Mean-Square loss), Ifconfig.num_labels > 1
a classification loss is computed (Cross-Entropy).
Returns
transformers.modeling_outputs.SequenceClassifierOutput or tuple(torch.FloatTensor)
A transformers.modeling_outputs.SequenceClassifierOutput 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 (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, if the model has an embedding layer, + 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 optional 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.
The CanineForSequenceClassification forward method, overrides the __call__
special method.
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.
Example of single-label classification:
>>> import torch
>>> from transformers import AutoTokenizer, CanineForSequenceClassification
>>> tokenizer = AutoTokenizer.from_pretrained("google/canine-s")
>>> model = CanineForSequenceClassification.from_pretrained("google/canine-s")
>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
>>> with torch.no_grad():
... logits = model(**inputs).logits
>>> predicted_class_id = logits.argmax().item()
>>> # To train a model on `num_labels` classes, you can pass `num_labels=num_labels` to `.from_pretrained(...)`
>>> num_labels = len(model.config.id2label)
>>> model = CanineForSequenceClassification.from_pretrained("google/canine-s", num_labels=num_labels)
>>> labels = torch.tensor([1])
>>> loss = model(**inputs, labels=labels).loss
Example of multi-label classification:
>>> import torch
>>> from transformers import AutoTokenizer, CanineForSequenceClassification
>>> tokenizer = AutoTokenizer.from_pretrained("google/canine-s")
>>> model = CanineForSequenceClassification.from_pretrained("google/canine-s", problem_type="multi_label_classification")
>>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
>>> with torch.no_grad():
... logits = model(**inputs).logits
>>> predicted_class_ids = torch.arange(0, logits.shape[-1])[torch.sigmoid(logits).squeeze(dim=0) > 0.5]
>>> # To train a model on `num_labels` classes, you can pass `num_labels=num_labels` to `.from_pretrained(...)`
>>> num_labels = len(model.config.id2label)
>>> model = CanineForSequenceClassification.from_pretrained(
... "google/canine-s", num_labels=num_labels, problem_type="multi_label_classification"
... )
>>> labels = torch.sum(
... torch.nn.functional.one_hot(predicted_class_ids[None, :].clone(), num_classes=num_labels), dim=1
... ).to(torch.float)
>>> loss = model(**inputs, labels=labels).loss
CanineForMultipleChoice
class transformers.CanineForMultipleChoice
< source >( config )
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 the from_pretrained() method to load the model weights.
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.
forward
< source >( input_ids: typing.Optional[torch.LongTensor] = None attention_mask: typing.Optional[torch.FloatTensor] = None token_type_ids: typing.Optional[torch.LongTensor] = None position_ids: typing.Optional[torch.LongTensor] = None head_mask: typing.Optional[torch.FloatTensor] = None inputs_embeds: typing.Optional[torch.FloatTensor] = None labels: typing.Optional[torch.LongTensor] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.modeling_outputs.MultipleChoiceModelOutput or tuple(torch.FloatTensor)
Parameters
- input_ids (
torch.LongTensor
of shape(batch_size, num_choices, sequence_length)
) — Indices of input sequence tokens in the vocabulary.Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and PreTrainedTokenizer.call() for details.
- attention_mask (
torch.FloatTensor
of shape(batch_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 shape(batch_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 shape(batch_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, num_choices, 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 a ModelOutput 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
transformers.modeling_outputs.MultipleChoiceModelOutput or tuple(torch.FloatTensor)
A transformers.modeling_outputs.MultipleChoiceModelOutput 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 (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, if the model has an embedding layer, + 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 optional 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.
The CanineForMultipleChoice forward method, overrides the __call__
special method.
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.
Example:
>>> from transformers import AutoTokenizer, CanineForMultipleChoice
>>> import torch
>>> tokenizer = AutoTokenizer.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
< source >( config )
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 the from_pretrained() method to load the model weights.
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.
forward
< source >( input_ids: typing.Optional[torch.LongTensor] = None attention_mask: typing.Optional[torch.FloatTensor] = None token_type_ids: typing.Optional[torch.LongTensor] = None position_ids: typing.Optional[torch.LongTensor] = None head_mask: typing.Optional[torch.FloatTensor] = None inputs_embeds: typing.Optional[torch.FloatTensor] = None labels: typing.Optional[torch.LongTensor] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.modeling_outputs.TokenClassifierOutput or tuple(torch.FloatTensor)
Parameters
- input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) — Indices of input sequence tokens in the vocabulary.Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and 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 a ModelOutput 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
transformers.modeling_outputs.TokenClassifierOutput or tuple(torch.FloatTensor)
A transformers.modeling_outputs.TokenClassifierOutput 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 (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, if the model has an embedding layer, + 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 optional 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.
The CanineForTokenClassification forward method, overrides the __call__
special method.
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.
Example:
>>> from transformers import AutoTokenizer, CanineForTokenClassification
>>> import torch
>>> tokenizer = AutoTokenizer.from_pretrained("google/canine-s")
>>> model = CanineForTokenClassification.from_pretrained("google/canine-s")
>>> inputs = tokenizer(
... "HuggingFace is a company based in Paris and New York", add_special_tokens=False, return_tensors="pt"
... )
>>> with torch.no_grad():
... logits = model(**inputs).logits
>>> predicted_token_class_ids = logits.argmax(-1)
>>> # Note that tokens are classified rather then input words which means that
>>> # there might be more predicted token classes than words.
>>> # Multiple token classes might account for the same word
>>> predicted_tokens_classes = [model.config.id2label[t.item()] for t in predicted_token_class_ids[0]]
>>> predicted_tokens_classes
CanineForQuestionAnswering
class transformers.CanineForQuestionAnswering
< source >( config )
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 the from_pretrained() method to load the model weights.
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.
forward
< source >( input_ids: typing.Optional[torch.LongTensor] = None attention_mask: typing.Optional[torch.FloatTensor] = None token_type_ids: typing.Optional[torch.LongTensor] = None position_ids: typing.Optional[torch.LongTensor] = None head_mask: typing.Optional[torch.FloatTensor] = None inputs_embeds: typing.Optional[torch.FloatTensor] = None start_positions: typing.Optional[torch.LongTensor] = None end_positions: typing.Optional[torch.LongTensor] = None output_attentions: typing.Optional[bool] = None output_hidden_states: typing.Optional[bool] = None return_dict: typing.Optional[bool] = None ) → transformers.modeling_outputs.QuestionAnsweringModelOutput or tuple(torch.FloatTensor)
Parameters
- input_ids (
torch.LongTensor
of shape(batch_size, sequence_length)
) — Indices of input sequence tokens in the vocabulary.Indices can be obtained using AutoTokenizer. See PreTrainedTokenizer.encode() and 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 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
transformers.modeling_outputs.QuestionAnsweringModelOutput or tuple(torch.FloatTensor)
A transformers.modeling_outputs.QuestionAnsweringModelOutput 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 (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, if the model has an embedding layer, + 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 optional 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.
The CanineForQuestionAnswering forward method, overrides the __call__
special method.
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.
Example:
>>> from transformers import AutoTokenizer, CanineForQuestionAnswering
>>> import torch
>>> tokenizer = AutoTokenizer.from_pretrained("Splend1dchan/canine-c-squad")
>>> model = CanineForQuestionAnswering.from_pretrained("Splend1dchan/canine-c-squad")
>>> question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet"
>>> inputs = tokenizer(question, text, return_tensors="pt")
>>> with torch.no_grad():
... outputs = model(**inputs)
>>> answer_start_index = outputs.start_logits.argmax()
>>> answer_end_index = outputs.end_logits.argmax()
>>> predict_answer_tokens = inputs.input_ids[0, answer_start_index : answer_end_index + 1]
>>> tokenizer.decode(predict_answer_tokens, skip_special_tokens=True)
'nice puppet'
>>> # target is "nice puppet"
>>> target_start_index = torch.tensor([14])
>>> target_end_index = torch.tensor([15])
>>> outputs = model(**inputs, start_positions=target_start_index, end_positions=target_end_index)
>>> loss = outputs.loss
>>> round(loss.item(), 2)
8.81