modeling

class TinyBertModel(vocab_size, 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=16, initializer_range=0.02, pad_token_id=0, fit_size=768)[source]

Bases: paddlenlp.transformers.tinybert.modeling.TinyBertPretrainedModel

The bare TinyBERT 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 TinyBertModel. Defines the number of different tokens that can be represented by the inputs_ids passed when calling TinyBertModel.

  • hidden_size (int, optional) – Dimensionality of the embedding layer, encoder layers 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. The dimensionality of position encoding is the dimensionality of the sequence in TinyBertModel. Defaults to 512.

  • type_vocab_size (int, optional) – The vocabulary size of token_type_ids passed when calling TinyBertModel. Defaults to 16.

  • 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 TinyBertPretrainedModel.init_weights() for how weights are initialized in TinyBertModel.

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

  • fit_size (int, optional) – Dimensionality of the output layer of fit_dense(s), which is the hidden size of the teacher model. fit_dense(s) means a hidden states’ transformation from student to teacher. fit_dense(s) will be generated when bert model is distilled during the training, and will not be generated during the prediction process. fit_denses is used in v2 models and it has num_hidden_layers+1 layers. fit_dense is used in other pretraining models and it has one linear layer. Defaults to 768.

get_input_embeddings() → paddle.nn.layer.common.Embedding[source]

get input embedding of TinyBert Pretrained Model

Returns

the input embedding of tiny bert

Return type

nn.Embedding

set_input_embeddings(embedding: paddle.nn.layer.common.Embedding) → None[source]

set the input embedding with the new embedding value

Parameters

embedding (nn.Embedding) – the new embedding value

forward(input_ids: Optional[paddle.Tensor] = None, token_type_ids: Optional[paddle.Tensor] = None, position_ids: Optional[paddle.Tensor] = None, attention_mask: Optional[paddle.Tensor] = None, inputs_embeds: Optional[paddle.Tensor] = None, past_key_values: Optional[Tuple[Tuple[paddle.Tensor]]] = None, use_cache: Optional[bool] = None, output_hidden_states: Optional[bool] = None, output_attentions: Optional[bool] = None, return_dict: Optional[bool] = None)[source]

The TinyBertModel 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 to some unwanted positions, usually the paddings or the subsequent positions. Its data type can be int, float and bool. When the data type is bool, the masked tokens have False values and the others have True values. When the data type is int, the masked tokens have 0 values and the others have 1 values. When the data type is float, the masked tokens have -INF values and the others have 0 values. It is a tensor with shape broadcasted to [batch_size, num_attention_heads, sequence_length, sequence_length]. For example, its shape can be [batch_size, sequence_length], [batch_size, sequence_length, sequence_length], [batch_size, num_attention_heads, sequence_length, sequence_length]. Defaults to None, which means nothing needed to be prevented attention to.

  • inputs_embeds (Tensor, optional) – If you want to control how to convert inputs_ids indices into associated vectors, you can pass an embedded representation directly instead of passing inputs_ids.

  • past_key_values (tuple(tuple(Tensor)), optional) – The length of tuple equals to the number of layers, and each inner tuple haves 4 tensors of shape (batch_size, num_heads, sequence_length - 1, embed_size_per_head)) which contains precomputed key and value hidden states of the attention blocks. If past_key_values are used, the user can optionally input only the last input_ids (those that don’t have their past key value states given to this model) of shape (batch_size, 1) instead of all input_ids of shape (batch_size, sequence_length).

  • use_cache (bool, optional) – If set to True, past_key_values key value states are returned. Defaults to None.

  • output_hidden_states (bool, optional) – Whether to return the hidden states of all layers. Defaults to False.

  • output_attentions (bool, optional) – Whether to return the attentions tensors of all attention layers. Defaults to False.

  • return_dict (bool, optional) – Whether to return a ModelOutput object. If False, the output will be a tuple of tensors. Defaults to False.

Returns

An instance of BaseModelOutputWithPoolingAndCrossAttentions if return_dict=True. Otherwise it returns a tuple of tensors corresponding to ordered and not None (depending on the input arguments) fields of BaseModelOutputWithPoolingAndCrossAttentions.

tuple: Returns tuple (encoder_output, pooled_output).

With the fields:

  • encoder_output (Tensor):

    Sequence of hidden-states at the last layer of the model. It’s data type should be float32 and its shape is [batch_size, sequence_length, hidden_size].

  • pooled_output (Tensor):

    The output of first token ([CLS]) in sequence. We “pool” the model by simply taking the hidden state corresponding to the first token. Its data type should be float32 and its shape is [batch_size, hidden_size].

Example

import paddle
from paddlenlp.transformers import TinyBertModel, TinyBertTokenizer

tokenizer = TinyBertTokenizer.from_pretrained('tinybert-4l-312d')
model = TinyBertModel.from_pretrained('tinybert-4l-312d')

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

Bases: paddlenlp.transformers.model_utils.PretrainedModel

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

init_weights(layer)[source]

Initialization hook

base_model_class

alias of paddlenlp.transformers.tinybert.modeling.TinyBertModel

class TinyBertForPretraining(tinybert)[source]

Bases: paddlenlp.transformers.tinybert.modeling.TinyBertPretrainedModel

TinyBert Model with pretraining tasks on top.

Parameters

tinybert (TinyBertModel) – An instance of TinyBertModel.

forward(input_ids: Optional[paddle.Tensor] = None, token_type_ids: Optional[paddle.Tensor] = None, position_ids: Optional[paddle.Tensor] = None, attention_mask: Optional[paddle.Tensor] = None, inputs_embeds: Optional[paddle.Tensor] = None, output_hidden_states: Optional[bool] = None, output_attentions: Optional[bool] = None, return_dict: Optional[bool] = None)[source]

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

Parameters
Returns

Returns tensor sequence_output, sequence of hidden-states at the last layer of the model. It’s data type should be float32 and its shape is [batch_size, sequence_length, hidden_size].

Return type

Tensor

Example

import paddle
from paddlenlp.transformers.tinybert.modeling import TinyBertForPretraining
from paddlenlp.transformers.tinybert.tokenizer import TinyBertTokenizer

tokenizer = TinyBertTokenizer.from_pretrained('tinybert-4l-312d')
model = TinyBertForPretraining.from_pretrained('tinybert-4l-312d')

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

logits = outputs[0]
class TinyBertForSequenceClassification(tinybert, num_classes=2, dropout=None)[source]

Bases: paddlenlp.transformers.tinybert.modeling.TinyBertPretrainedModel

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

Parameters

  • tinybert (TinyBertModel) – An instance of TinyBertModel.

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

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

forward(input_ids: Optional[paddle.Tensor] = None, token_type_ids: Optional[paddle.Tensor] = None, position_ids: Optional[paddle.Tensor] = None, attention_mask: Optional[paddle.Tensor] = None, labels: Optional[paddle.Tensor] = None, inputs_embeds: Optional[paddle.Tensor] = None, output_hidden_states: Optional[bool] = None, output_attentions: Optional[bool] = None, return_dict: Optional[bool] = None)[source]

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

Parameters

  • input_ids (Tensor) – See TinyBertModel.

  • token_type_ids (Tensor, optional) – See TinyBertModel.

  • position_ids (Tensor, optional) – See TinyBertModel.

  • attention_mask_list (list, optional) – See TinyBertModel.

  • labels (Tensor of shape (batch_size,), optional) – Labels for computing the sequence classification/regression loss. Indices should be in [0, ..., num_classes - 1]. If num_classes == 1 a regression loss is computed (Mean-Square loss), If num_classes > 1 a classification loss is computed (Cross-Entropy).

  • output_hidden_states (bool, optional) – Whether to return the hidden states of all layers. Defaults to False.

  • output_attentions (bool, optional) – Whether to return the attentions tensors of all attention layers. Defaults to False.

  • return_dict (bool, optional) – Whether to return a SequenceClassifierOutput object. If False, the output will be a tuple of tensors. Defaults to False.

Returns

An instance of SequenceClassifierOutput if return_dict=True. Otherwise it returns a tuple of tensors corresponding to ordered and not None (depending on the input arguments) fields of SequenceClassifierOutput.

Example

import paddle
from paddlenlp.transformers.tinybert.modeling import TinyBertForSequenceClassification
from paddlenlp.transformers.tinybert.tokenizer import TinyBertTokenizer

tokenizer = TinyBertTokenizer.from_pretrained('tinybert-4l-312d')
model = TinyBertForSequenceClassification.from_pretrained('tinybert-4l-312d')

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

logits = outputs[0]
class TinyBertForQuestionAnswering(tinybert)[source]

Bases: paddlenlp.transformers.tinybert.modeling.TinyBertPretrainedModel

TinyBert 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

tinybert (TinyBertModel) – An instance of TinyBertModel.

forward(input_ids: Optional[paddle.Tensor] = None, token_type_ids: Optional[paddle.Tensor] = None, position_ids: Optional[paddle.Tensor] = None, attention_mask: Optional[paddle.Tensor] = None, inputs_embeds: Optional[paddle.Tensor] = None, start_positions: Optional[paddle.Tensor] = None, end_positions: Optional[paddle.Tensor] = None, output_hidden_states: Optional[bool] = None, output_attentions: Optional[bool] = None, return_dict: Optional[bool] = None)[source]
Parameters

  • input_ids (Tensor) – See TinyBertModel.

  • token_type_ids (Tensor, optional) – See TinyBertModel.

  • position_ids (Tensor, optional) – See TinyBertModel.

  • attention_mask (Tensor, optional) – See TinyBertModel.

  • start_positions (Tensor 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 (Tensor 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.

  • output_hidden_states (bool, optional) – Whether to return the hidden states of all layers. Defaults to False.

  • output_attentions (bool, optional) – Whether to return the attentions tensors of all attention layers. Defaults to False.

  • return_dict (bool, optional) – Whether to return a QuestionAnsweringModelOutput object. If False, the output will be a tuple of tensors. Defaults to False.

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 TinyBertForQuestionAnswering, TinyBertTokenizer

tokenizer = TinyBertTokenizer.from_pretrained('tinybert-6l-768d-zh')
model = TinyBertForQuestionAnswering.from_pretrained('tinybert-6l-768d-zh')

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

Bases: paddlenlp.transformers.tinybert.modeling.TinyBertPretrainedModel

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

Parameters

  • tinybert (TinyBertModel) – An instance of TinyBertModel.

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

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

forward(input_ids: Optional[paddle.Tensor] = None, token_type_ids: Optional[paddle.Tensor] = None, position_ids: Optional[paddle.Tensor] = None, attention_mask: Optional[paddle.Tensor] = None, inputs_embeds: Optional[paddle.Tensor] = None, labels: Optional[paddle.Tensor] = None, output_hidden_states: Optional[bool] = None, output_attentions: Optional[bool] = None, return_dict: Optional[bool] = None)[source]

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

Parameters

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

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

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

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

  • labels (Tensor of shape (batch_size, ), optional) – Labels for computing the multiple choice classification loss. Indices should be in [0, ..., num_choices-1] where num_choices is the size of the second dimension of the input tensors. (See input_ids above)

  • output_hidden_states (bool, optional) – Whether to return the hidden states of all layers. Defaults to False.

  • output_attentions (bool, optional) – Whether to return the attentions tensors of all attention layers. Defaults to False.

  • return_dict (bool, optional) – Whether to return a MultipleChoiceModelOutput object. If False, the output will be a tuple of tensors. Defaults to False.

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