modeling

class MegatronBertModel(vocab_size=29056, hidden_size=1024, pad_token_id=0, type_vocab_size=2, hidden_act='gelu', attention_probs_dropout_prob=0.1, num_attention_heads=16, num_hidden_layers=24, max_position_embeddings=512, hidden_dropout_prob=0.1, intermediate_size=4096, position_embedding_type='absolute', initializer_range=0.02)[source]

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

The bare MegatronBert 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 MegatronBertModel. 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 MegatronBert.

  • hidden_size (int, optional) – Dimensionality of the encoder layer and pooler layer. Defaults to 1024.

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

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

  • 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".

  • 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.

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

  • num_hidden_layers (int, optional) – Number of hidden layers in the Transformer encoder. Defaults to 24.

  • 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.

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

  • 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 4096.

  • position_embedding_type (str, optional) – Type of position embedding. Defaults to “absolute”

  • 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 MegatronBertPretrainedModel.init_weights() for how weights are initialized in MegatronBertModel.

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=None, token_type_ids=None, position_ids=None, attention_mask=None)[source]

The MegatronBertModel 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 tuple (sequence_output, pooled_output).

With the fields:

  • sequence_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].

Return type

tuple

Example

import paddle
from paddlenlp.transformers import MegatronBertModel, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertModel.from_pretrained('megatronbert-uncased')

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

Bases: paddlenlp.transformers.model_utils.PretrainedModel

An abstract class for pretrained MegatronBert models. It provides RoBerta 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(layer)[source]

Initialization hook

base_model_class

alias of paddlenlp.transformers.megatronbert.modeling.MegatronBertModel

class MegatronBertForQuestionAnswering(megatronbert)[source]

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

MegatronBert Model with question answering tasks.

Parameters

megatronbert (MegatronBertModel) – An instance of MegatronBertModel.

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

The MegatronBertForQuestionAnswering 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 MegatronBertForQuestionAnswering, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertForQuestionAnswering.from_pretrained('megatronbert-uncased')

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 MegatronBertForSequenceClassification(megatronbert, num_labels)[source]

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

MegatronBert Model with sequence classification tasks.

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

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

Parameters
Returns

Returns tensor logits, a tensor of the sequence classification logits.

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import MegatronBertForSequenceClassification, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertForSequenceClassification.from_pretrained('megatronbert-uncased', num_labels=2)

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

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

MegatronBert Model with a next sentence prediction (classification) head on top.

Parameters

megatronbert (MegatronBertModel) – An instance of MegatronBertModel.

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

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

Parameters
Returns

Returns Tensor seq_relationship_scores. The scores of next sentence prediction.

Its data type should be float32 and its shape is [batch_size, 2].

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import MegatronBertForNextSentencePrediction, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertForNextSentencePrediction.from_pretrained('megatronbert-uncased')

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

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

MegatronBert Model with a causal masked language modeling head on top.

Parameters

megatronbert (MegatronBertModel) – An instance of MegatronBertModel.

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

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

Parameters
Returns

Returns Tensor prediction_scores. The scores of masked token prediction.

Its data type should be float32. If masked_positions is None, its shape is [batch_size, sequence_length, vocab_size]. Otherwise, its shape is [batch_size, mask_token_num, vocab_size].

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import MegatronBertForCausalLM, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertForCausalLM.from_pretrained('megatronbert-uncased')

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 MegatronBertForPreTraining(megatronbert)[source]

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

Megatronbert Model with pretraining tasks on top.

Parameters

megatronbert (MegatronBertModel) – An instance of MegatronBertModel.

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

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

Parameters
Returns

Returns tuple (prediction_scores, seq_relationship_score).

With the fields:

  • prediction_scores (Tensor):

    The scores of masked token prediction. Its data type should be float32. If masked_positions is None, its shape is [batch_size, sequence_length, vocab_size]. Otherwise, its shape is [batch_size, mask_token_num, vocab_size].

  • seq_relationship_score (Tensor):

    The scores of next sentence prediction. Its data type should be float32 and its shape is [batch_size, 2].

Return type

tuple

Example

import paddle
from paddlenlp.transformers import MegatronBertForPreTraining, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertForPreTraining.from_pretrained('megatronbert-uncased')

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

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

MegatronBert Model with a masked language modeling head on top.

Parameters

megatronbert (MegatronBertModel) – An instance of MegatronBertModel.

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

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

Parameters
Returns

Returns Tensor prediction_scores. The scores of masked token prediction.

Its data type should be float32. If masked_positions is None, its shape is [batch_size, sequence_length, vocab_size]. Otherwise, its shape is [batch_size, mask_token_num, vocab_size].

Return type

Tensor

Example

import paddle
from paddlenlp.transformers import MegatronBertForMaskedLM, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertForMaskedLM.from_pretrained('megatronbert-uncased')

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 MegatronBertForMultipleChoice(megatronbert)[source]

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

MegatronBert Model with a multiple choice classification head on top.

Parameters

megatronbert (MegatronBertModel) – An instance of MegatronBertModel.

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

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

Parameters
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 MegatronBertForMultipleChoice, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertForNextSentencePrediction.from_pretrained('megatronbert-uncased')

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

Bases: paddlenlp.transformers.megatronbert.modeling.MegatronBertPretrainedModel

MegatronBert Model with a token classification head on top.

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

The MegatronBertForTokenClassification 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 MegatronBertForTokenClassification, MegatronBertTokenizer

tokenizer = MegatronBertTokenizer.from_pretrained('megatronbert-uncased')
model = MegatronBertForTokenClassification.from_pretrained('megatronbert-uncased', num_labels=2)

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