modeling

class ConvBertModel(vocab_size, embedding_size=768, 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=512, type_vocab_size=2, initializer_range=0.02, pad_token_id=0, conv_kernel_size=9, head_ratio=2, num_groups=1)[source]

Bases: paddlenlp.transformers.convbert.modeling.ConvBertPretrainedModel

The bare ConvBert Model transformer outputting raw hidden-states.

This model inherits from PretrainedModel. Refer to the superclass documentation for the generic methods.

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

Parameters
  • vocab_size (int) – Vocabulary size of inputs_ids in ConvBertModel. Also is the vocab size of token embedding matrix. Defines the number of different tokens that can be represented by the inputs_ids passed when calling ConvBertModel.

  • embedding_size (int, optional) – Dimensionality of the embedding layer. Defaults to 768.

  • hidden_size (int, optional) – Dimensionality of the encoder layer and pooler layer. Defaults to 768.

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

  • num_attention_heads (int, optional) – Number of attention heads for each attention layer in the Transformer encoder. Defaults to 12.

  • intermediate_size (int, optional) – Dimensionality of the feed-forward (ff) layer in the encoder. Input tensors to ff layers are firstly projected from hidden_size to intermediate_size, and then projected back to hidden_size. Typically intermediate_size is larger than hidden_size. Defaults to 3072.

  • hidden_act (str, optional) – The non-linear activation function in the feed-forward layer. "gelu", "relu" and any other paddle supported activation functions are supported. Defaults to "gelu".

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

  • attention_probs_dropout_prob (float, optional) – The dropout probability used in MultiHeadAttention in all encoder layers to drop some attention target. Defaults to 0.1.

  • max_position_embeddings (int, optional) – The maximum value of the dimensionality of position encoding, which dictates the maximum supported length of an input sequence. Defaults to 512.

  • type_vocab_size (int, optional) – The vocabulary size of token_type_ids. Defaults to 2.

  • initializer_range (float, optional) –

    The standard deviation of the normal initializer. Defaults to 0.02.

    Note

    A normal_initializer initializes weight matrices as normal distributions. See ConvBertPretrainedModel.init_weights() for how weights are initialized in ConvBertModel.

  • pad_token_id (int, optional) – The index of padding token in the token vocabulary. Defaults to 0.

  • conv_kernel_size (int, optional) – The size of the convolutional kernel. Defaults to 9.

  • head_ratio (int, optional) – Ratio gamma to reduce the number of attention heads. Defaults to 2.

  • num_groups (int, optional) – The number of groups for grouped linear layers for ConvBert model. Defaults to 1.

get_input_embeddings()[source]

get input embedding of model

Returns

embedding of model

Return type

nn.Embedding

set_input_embeddings(value)[source]

set new input embedding for model

Parameters

value (Embedding) – the new embedding of model

Raises

NotImplementedError – Model has not implement set_input_embeddings method

forward(input_ids, token_type_ids=None, position_ids=None, attention_mask=None)[source]

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

Parameters
  • input_ids (Tensor) – Indices of input sequence tokens in the vocabulary. They are numerical representations of tokens that build the input sequence. Its data type should be int64 and it has a shape of [batch_size, sequence_length].

  • token_type_ids (Tensor, optional) –

    Segment token indices to indicate different portions of the inputs. Selected in the range [0, type_vocab_size - 1]. If type_vocab_size is 2, which means the inputs have two portions. Indices can either be 0 or 1:

    • 0 corresponds to a sentence A token,

    • 1 corresponds to a sentence B token.

    Its data type should be int64 and it has a shape of [batch_size, sequence_length]. Defaults to None, which means we don’t add segment embeddings.

  • position_ids (Tensor, optional) – Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, max_position_embeddings - 1]. Shape as (batch_size, num_tokens) and dtype as int64. Defaults to None.

  • attention_mask (Tensor, optional) –

    Mask used in multi-head attention to avoid performing attention on to some unwanted positions, usually the paddings or the subsequent positions. Its data type can be int, float and bool. If its data type is int, the values should be either 0 or 1.

    • 1 for tokens that not masked,

    • 0 for tokens that masked.

    It is a tensor with shape broadcasted to [batch_size, num_attention_heads, sequence_length, sequence_length]. Defaults to None, which means nothing needed to be prevented attention to.

Returns

Returns Tensor sequence_output, sequence of hidden-states at the last layer of the model. Shape as [batch_size, sequence_length, hidden_size] and dtype as float32.

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import ConvBertModel, ConvBertTokenizer

tokenizer = ConvBertTokenizer.from_pretrained('convbert-base')
model = ConvBertModel.from_pretrained('convbert-base')

inputs = tokenizer("Welcome to use PaddlePaddle and PaddleNLP!")
inputs = {k:paddle.to_tensor([v]) for (k, v) in inputs.items()}
output = model(**inputs)
class ConvBertPretrainedModel(*args, **kwargs)[source]

Bases: paddlenlp.transformers.model_utils.PretrainedModel

An abstract class for pretrained ConvBert models. It provides ConvBert related model_config_file, pretrained_init_configuration, resource_files_names, pretrained_resource_files_map, base_model_prefix for downloading and loading pretrained models. See PretrainedModel for more details.

init_weights()[source]

Initializes and tie weights if needed.

tie_weights()[source]

Tie the weights between the input embeddings and the output embeddings.

base_model_class

alias of paddlenlp.transformers.convbert.modeling.ConvBertModel

class ConvBertForTotalPretraining(generator, discriminator)[source]

Bases: paddlenlp.transformers.convbert.modeling.ConvBertPretrainedModel

Combine generator with discriminator for Replaced Token Detection (RTD) pretraining.

get_input_embeddings()[source]

get input embedding of model

Returns

embedding of model

Return type

nn.Embedding

get_output_embeddings()[source]

To be overwrited for models with output embeddings

Returns

the otuput embedding of model

Return type

Optional[Embedding]

get_discriminator_inputs(inputs, raw_inputs, gen_logits, gen_labels, use_softmax_sample)[source]

Sample from the generator to create discriminator input.

forward(input_ids=None, token_type_ids=None, position_ids=None, attention_mask=None, raw_input_ids=None, gen_labels=None)[source]
Parameters
  • input_ids (Tensor) – See ConvBertModel.

  • token_type_ids (Tensor, optional) – See ConvBertModel.

  • position_ids (Tensor, optional) – See ConvBertModel.

  • attention_mask (Tensor, optional) – See ConvBertModel.

  • raw_input_ids (Tensor, optional) – The raw input_ids. Its data type should be int64 and it has a shape of [batch_size, sequence_length].

  • gen_labels (Tensor, optional) – The generator labels. Its data type should be int64 and it has a shape of [batch_size, sequence_length].

Returns

Returns tuple (gen_logits, disc_logits, disc_labels, attention_mask).

With the fields:

  • gen_logits (Tensor):

    a tensor of the generator prediction logits. Shape as [batch_size, sequence_length, vocab_size] and dtype as float32.

  • disc_logits (Tensor):

    a tensor of the discriminator prediction logits. Shape as [batch_size, sequence_length] and dtype as float32.

  • disc_labels (Tensor):

    a tensor of the discriminator prediction labels. Shape as [batch_size, sequence_length] and dtype as int64.

  • attention_mask (Tensor):

    See ConvBertModel.

Return type

tuple

class ConvBertDiscriminator(convbert)[source]

Bases: paddlenlp.transformers.convbert.modeling.ConvBertPretrainedModel

ConvBert Model with a discriminator prediction head on top.

Parameters

convbert (ConvBertModel) – An instance of ConvBertModel.

forward(input_ids, token_type_ids=None, position_ids=None, attention_mask=None)[source]

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

Parameters
  • input_ids (Tensor) – Indices of input sequence tokens in the vocabulary. They are numerical representations of tokens that build the input sequence. Its data type should be int64 and it has a shape of [batch_size, sequence_length].

  • token_type_ids (Tensor, optional) –

    Segment token indices to indicate different portions of the inputs. Selected in the range [0, type_vocab_size - 1]. If type_vocab_size is 2, which means the inputs have two portions. Indices can either be 0 or 1:

    • 0 corresponds to a sentence A token,

    • 1 corresponds to a sentence B token.

    Its data type should be int64 and it has a shape of [batch_size, sequence_length]. Defaults to None, which means we don’t add segment embeddings.

  • position_ids (Tensor, optional) – Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, max_position_embeddings - 1]. Shape as (batch_size, num_tokens) and dtype as int64. Defaults to None.

  • attention_mask (Tensor, optional) –

    Mask used in multi-head attention to avoid performing attention on to some unwanted positions, usually the paddings or the subsequent positions. Its data type can be int, float and bool. If its data type is int, the values should be either 0 or 1.

    • 1 for tokens that not masked,

    • 0 for tokens that masked.

    It is a tensor with shape broadcasted to [batch_size, num_attention_heads, sequence_length, sequence_length]. Defaults to None, which means nothing needed to be prevented attention to.

Returns

Returns tensor logits, a tensor of the discriminator prediction logits. Shape as [batch_size, sequence_length] and dtype as float32.

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import ConvBertDiscriminatorPredictions, ConvBertTokenizer

tokenizer = ConvBertTokenizer.from_pretrained('convbert-base')
model = ConvBertDiscriminatorPredictions.from_pretrained('convbert-base')

inputs = tokenizer("Welcome to use PaddlePaddle and PaddleNLP!")
inputs = {k:paddle.to_tensor([v]) for (k, v) in inputs.items()}
logits = model(**inputs)
class ConvBertGenerator(convbert)[source]

Bases: paddlenlp.transformers.convbert.modeling.ConvBertPretrainedModel

ConvBert Model with a generator prediction head on top.

Parameters

convbert (ConvBertModel) – An instance of ConvBertModel.

get_input_embeddings()[source]

get input embedding of model

Returns

embedding of model

Return type

nn.Embedding

forward(input_ids=None, token_type_ids=None, position_ids=None, attention_mask=None)[source]

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

Parameters
  • input_ids (Tensor) – Indices of input sequence tokens in the vocabulary. They are numerical representations of tokens that build the input sequence. Its data type should be int64 and it has a shape of [batch_size, sequence_length].

  • token_type_ids (Tensor, optional) –

    Segment token indices to indicate different portions of the inputs. Selected in the range [0, type_vocab_size - 1]. If type_vocab_size is 2, which means the inputs have two portions. Indices can either be 0 or 1:

    • 0 corresponds to a sentence A token,

    • 1 corresponds to a sentence B token.

    Its data type should be int64 and it has a shape of [batch_size, sequence_length]. Defaults to None, which means we don’t add segment embeddings.

  • position_ids (Tensor, optional) – Indices of positions of each input sequence tokens in the position embeddings. Selected in the range [0, max_position_embeddings - 1]. Shape as (batch_size, num_tokens) and dtype as int64. Defaults to None.

  • attention_mask (Tensor, optional) –

    Mask used in multi-head attention to avoid performing attention on to some unwanted positions, usually the paddings or the subsequent positions. Its data type can be int, float and bool. If its data type is int, the values should be either 0 or 1.

    • 1 for tokens that not masked,

    • 0 for tokens that masked.

    It is a tensor with shape broadcasted to [batch_size, num_attention_heads, sequence_length, sequence_length]. Defaults to None, which means nothing needed to be prevented attention to.

Returns

Returns tensor prediction_scores, a tensor of the generator prediction scores. Shape as [batch_size, sequence_length, vocab_size] and dtype as float32.

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import ConvBertGenerator, ConvBertTokenizer

tokenizer = ConvBertTokenizer.from_pretrained('convbert-base')
model = ConvBertGenerator.from_pretrained('convbert-base')

inputs = tokenizer("Welcome to use PaddlePaddle and PaddleNLP!")
inputs = {k:paddle.to_tensor([v]) for (k, v) in inputs.items()}
prediction_scores = model(**inputs)
class ConvBertClassificationHead(hidden_size, hidden_dropout_prob, num_classes)[source]

Bases: paddle.fluid.dygraph.layers.Layer

ConvBert head for sentence-level classification tasks.

forward(features, **kwargs)[source]

Defines the computation performed at every call. Should be overridden by all subclasses.

Parameters
  • *inputs (tuple) – unpacked tuple arguments

  • **kwargs (dict) – unpacked dict arguments

class ConvBertForSequenceClassification(convbert, num_classes=2, dropout=None)[source]

Bases: paddlenlp.transformers.convbert.modeling.ConvBertPretrainedModel

ConvBert Model with a linear layer on top of the output layer, designed for sequence classification/regression tasks like GLUE tasks.

Parameters
  • convbert (ConvBertModel) – An instance of ConvBertModel.

  • num_classes (int, optional) – The number of classes. Defaults to 2.

  • dropout (float, optional) – The dropout probability for output of ConvBert. If None, use the same value as hidden_dropout_prob of ConvBertModel instance convbert. Defaults to None.

forward(input_ids=None, token_type_ids=None, position_ids=None, attention_mask=None)[source]

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

Parameters
Returns

Returns tensor logits, a tensor of the input text classification logits. Shape as [batch_size, num_classes] and dtype as float32.

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import ConvBertForSequenceClassification, ConvBertTokenizer

tokenizer = ConvBertTokenizer.from_pretrained('convbert-base')
model = ConvBertForSequenceClassification.from_pretrained('convbert-base')

inputs = tokenizer("Welcome to use PaddlePaddle and PaddleNLP!")
inputs = {k:paddle.to_tensor([v]) for (k, v) in inputs.items()}
logits = model(**inputs)
class ConvBertForTokenClassification(convbert, num_classes=2, dropout=None)[source]

Bases: paddlenlp.transformers.convbert.modeling.ConvBertPretrainedModel

ConvBert Model with a linear layer on top of the hidden-states output layer, designed for token classification tasks like NER tasks.

Parameters
  • convbert (ConvBertModel) – An instance of ConvBertModel.

  • num_classes (int, optional) – The number of classes. Defaults to 2.

  • dropout (float, optional) – The dropout probability for output of ConvBert. If None, use the same value as hidden_dropout_prob of ConvBertModel instance convbert. Defaults to None.

forward(input_ids=None, token_type_ids=None, position_ids=None, attention_mask=None)[source]

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

Parameters
Returns

Returns tensor logits, a tensor of the input token classification logits. Shape as [batch_size, sequence_length, num_classes] and dtype as float32.

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import ConvBertForTokenClassification, ConvBertTokenizer

tokenizer = ConvBertTokenizer.from_pretrained('convbert-base')
model = ConvBertForTokenClassification.from_pretrained('convbert-base')

inputs = tokenizer("Welcome to use PaddlePaddle and PaddleNLP!")
inputs = {k:paddle.to_tensor([v]) for (k, v) in inputs.items()}
logits = model(**inputs)
class ConvBertPretrainingCriterion(vocab_size, gen_weight, disc_weight)[source]

Bases: paddle.fluid.dygraph.layers.Layer

Parameters
  • vocab_size (int) – Vocabulary size of inputs_ids in ConvBertModel. Defines the number of different tokens that can be represented by the inputs_ids passed when calling ConvBertModel.

  • gen_weight (float) – This is the generator weight.

  • disc_weight (float) – This is the discriminator weight.

forward(generator_prediction_scores, discriminator_prediction_scores, generator_labels, discriminator_labels, attention_mask)[source]

Defines the computation performed at every call. Should be overridden by all subclasses.

Parameters
  • *inputs (tuple) – unpacked tuple arguments

  • **kwargs (dict) – unpacked dict arguments

class ConvBertForQuestionAnswering(convbert)[source]

Bases: paddlenlp.transformers.convbert.modeling.ConvBertPretrainedModel

ConvBert Model with a linear layer on top of the hidden-states output to compute span_start_logits and span_end_logits, designed for question-answering tasks like SQuAD.

Parameters

convbert (ConvBertModel) – An instance of ConvBertModel.

forward(input_ids, token_type_ids=None, position_ids=None, attention_mask=None)[source]

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

Parameters
Returns

Returns tuple (start_logits, end_logits).

With the fields:

  • start_logits (Tensor):

    A tensor of the input token classification logits, indicates the start position of the labelled span. Its data type should be float32 and its shape is [batch_size, sequence_length].

  • end_logits (Tensor):

    A tensor of the input token classification logits, indicates the end position of the labelled span. Its data type should be float32 and its shape is [batch_size, sequence_length].

Return type

tuple

Example

import paddle
from paddlenlp.transformers import ConvBertForQuestionAnswering, ConvBertTokenizer

tokenizer = ConvBertTokenizer.from_pretrained('convbert-base')
model = ConvBertForQuestionAnswering.from_pretrained('convbert-base')

inputs = tokenizer("Welcome to use PaddlePaddle and PaddleNLP!")
inputs = {k:paddle.to_tensor([v]) for (k, v) in inputs.items()}
outputs = model(**inputs)

start_logits = outputs[0]
end_logits  = outputs[1]
class ConvBertForMultipleChoice(convbert, num_choices=2, dropout=None)[source]

Bases: paddlenlp.transformers.convbert.modeling.ConvBertPretrainedModel

ConvBert Model with a linear layer on top of the hidden-states output layer, designed for multiple choice tasks like RocStories/SWAG tasks .

Parameters
  • convbert (ConvBertModel) – An instance of ConvBertModel.

  • num_choices (int, optional) – The number of choices. Defaults to 2.

  • dropout (float, optional) – The dropout probability for output of ConvBert. If None, use the same value as hidden_dropout_prob of ConvBertModel instance convbert. Defaults to None.

forward(input_ids=None, token_type_ids=None, position_ids=None, attention_mask=None)[source]

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

Parameters
  • input_ids (Tensor) – See ConvBertModel and shape as [batch_size,num_choice, sequence_length].

  • token_type_ids (Tensor, optional) – See ConvBertModel and shape as [batch_size,num_choice, sequence_length].

  • position_ids (Tensor, optional) – See ConvBertModel and shape as [batch_size,num_choice, sequence_length].

  • attention_mask (list, optional) – See ConvBertModel and shape as [batch_size,num_choice, sequence_length].

Returns

Returns tensor reshaped_logits, a tensor of the multiple choice classification logits. Shape as [batch_size, num_choice] and dtype as float32.

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import ConvBertForMultipleChoice, ConvBertTokenizer

tokenizer = ConvBertTokenizer.from_pretrained('convbert-base')
model = ConvBertForMultipleChoice.from_pretrained('convbert-base')

inputs = tokenizer("Welcome to use PaddlePaddle and PaddleNLP!")
inputs = {k:paddle.to_tensor([v]) for (k, v) in inputs.items()}
logits = model(**inputs)
ConvBertForPretraining

alias of paddlenlp.transformers.convbert.modeling.ConvBertForTotalPretraining