HuggingFace_transformer/src/transformers/modeling_roberta.py

# coding=utf-8
# Copyright 2018 The Google AI Language Team Authors and The HuggingFace Inc. team.
# Copyright (c) 2018, NVIDIA CORPORATION.  All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""PyTorch RoBERTa model. """


import logging
import warnings

import torch
import torch.nn as nn
from torch.nn import CrossEntropyLoss, MSELoss

from .configuration_roberta import RobertaConfig
from .file_utils import add_code_sample_docstrings, add_start_docstrings, add_start_docstrings_to_callable
from .modeling_bert import BertEmbeddings, BertLayerNorm, BertModel, BertPreTrainedModel, gelu
from .modeling_outputs import (
    MaskedLMOutput,
    MultipleChoiceModelOutput,
    QuestionAnsweringModelOutput,
    SequenceClassifierOutput,
    TokenClassifierOutput,
)


logger = logging.getLogger(__name__)

_CONFIG_FOR_DOC = "RobertaConfig"
_TOKENIZER_FOR_DOC = "RobertaTokenizer"

ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST = [
    "roberta-base",
    "roberta-large",
    "roberta-large-mnli",
    "distilroberta-base",
    "roberta-base-openai-detector",
    "roberta-large-openai-detector",
    # See all RoBERTa models at https://huggingface.co/models?filter=roberta
]


class RobertaEmbeddings(BertEmbeddings):
    """
    Same as BertEmbeddings with a tiny tweak for positional embeddings indexing.
    """

    def __init__(self, config):
        super().__init__(config)
        self.padding_idx = config.pad_token_id
        self.word_embeddings = nn.Embedding(config.vocab_size, config.hidden_size, padding_idx=self.padding_idx)
        self.position_embeddings = nn.Embedding(
            config.max_position_embeddings, config.hidden_size, padding_idx=self.padding_idx
        )

    def forward(self, input_ids=None, token_type_ids=None, position_ids=None, inputs_embeds=None):
        if position_ids is None:
            if input_ids is not None:
                # Create the position ids from the input token ids. Any padded tokens remain padded.
                position_ids = create_position_ids_from_input_ids(input_ids, self.padding_idx).to(input_ids.device)
            else:
                position_ids = self.create_position_ids_from_inputs_embeds(inputs_embeds)

        return super().forward(
            input_ids, token_type_ids=token_type_ids, position_ids=position_ids, inputs_embeds=inputs_embeds
        )

    def create_position_ids_from_inputs_embeds(self, inputs_embeds):
        """ We are provided embeddings directly. We cannot infer which are padded so just generate
        sequential position ids.

        :param torch.Tensor inputs_embeds:
        :return torch.Tensor:
        """
        input_shape = inputs_embeds.size()[:-1]
        sequence_length = input_shape[1]

        position_ids = torch.arange(
            self.padding_idx + 1, sequence_length + self.padding_idx + 1, dtype=torch.long, device=inputs_embeds.device
        )
        return position_ids.unsqueeze(0).expand(input_shape)


ROBERTA_START_DOCSTRING = r"""

    This model is a PyTorch `torch.nn.Module <https://pytorch.org/docs/stable/nn.html#torch.nn.Module>`_ sub-class.
    Use it as a regular PyTorch Module and refer to the PyTorch documentation for all matter related to general
    usage and behavior.

    Parameters:
        config (:class:`~transformers.RobertaConfig`): 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 :meth:`~transformers.PreTrainedModel.from_pretrained` method to load the model weights.
"""

ROBERTA_INPUTS_DOCSTRING = r"""
    Args:
        input_ids (:obj:`torch.LongTensor` of shape :obj:`{0}`):
            Indices of input sequence tokens in the vocabulary.

            Indices can be obtained using :class:`transformers.RobertaTokenizer`.
            See :func:`transformers.PreTrainedTokenizer.encode` and
            :func:`transformers.PreTrainedTokenizer.__call__` for details.

            `What are input IDs? <../glossary.html#input-ids>`__
        attention_mask (:obj:`torch.FloatTensor` of shape :obj:`{0}`, `optional`, defaults to :obj:`None`):
            Mask to avoid performing attention on padding token indices.
            Mask values selected in ``[0, 1]``:
            ``1`` for tokens that are NOT MASKED, ``0`` for MASKED tokens.

            `What are attention masks? <../glossary.html#attention-mask>`__
        token_type_ids (:obj:`torch.LongTensor` of shape :obj:`{0}`, `optional`, defaults to :obj:`None`):
            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

            `What are token type IDs? <../glossary.html#token-type-ids>`_
        position_ids (:obj:`torch.LongTensor` of shape :obj:`{0}`, `optional`, defaults to :obj:`None`):
            Indices of positions of each input sequence tokens in the position embeddings.
            Selected in the range ``[0, config.max_position_embeddings - 1]``.

            `What are position IDs? <../glossary.html#position-ids>`_
        head_mask (:obj:`torch.FloatTensor` of shape :obj:`(num_heads,)` or :obj:`(num_layers, num_heads)`, `optional`, defaults to :obj:`None`):
            Mask to nullify selected heads of the self-attention modules.
            Mask values selected in ``[0, 1]``:
            :obj:`1` indicates the head is **not masked**, :obj:`0` indicates the head is **masked**.
        inputs_embeds (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, sequence_length, hidden_size)`, `optional`, defaults to :obj:`None`):
            Optionally, instead of passing :obj:`input_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 (:obj:`bool`, `optional`, defaults to :obj:`None`):
            If set to ``True``, the attentions tensors of all attention layers are returned. See ``attentions`` under returned tensors for more detail.
        output_hidden_states (:obj:`bool`, `optional`, defaults to :obj:`None`):
            If set to ``True``, the hidden states of all layers are returned. See ``hidden_states`` under returned tensors for more detail.
        return_tuple (:obj:`bool`, `optional`, defaults to :obj:`None`):
            If set to ``True``, the output of the model will be a plain tuple instead of a ``dataclass``.
"""


@add_start_docstrings(
    "The bare RoBERTa Model transformer outputting raw hidden-states without any specific head on top.",
    ROBERTA_START_DOCSTRING,
)
class RobertaModel(BertModel):
    """
    This class overrides :class:`~transformers.BertModel`. Please check the
    superclass for the appropriate documentation alongside usage examples.
    """

    config_class = RobertaConfig
    base_model_prefix = "roberta"

    def __init__(self, config):
        super().__init__(config)

        self.embeddings = RobertaEmbeddings(config)
        self.init_weights()

    def get_input_embeddings(self):
        return self.embeddings.word_embeddings

    def set_input_embeddings(self, value):
        self.embeddings.word_embeddings = value


@add_start_docstrings("""RoBERTa Model with a `language modeling` head on top. """, ROBERTA_START_DOCSTRING)
class RobertaForMaskedLM(BertPreTrainedModel):
    config_class = RobertaConfig
    base_model_prefix = "roberta"

    def __init__(self, config):
        super().__init__(config)

        self.roberta = RobertaModel(config)
        self.lm_head = RobertaLMHead(config)

        self.init_weights()

    def get_output_embeddings(self):
        return self.lm_head.decoder

    @add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("(batch_size, sequence_length)"))
    @add_code_sample_docstrings(
        tokenizer_class=_TOKENIZER_FOR_DOC,
        checkpoint="roberta-base",
        output_type=MaskedLMOutput,
        config_class=_CONFIG_FOR_DOC,
    )
    def forward(
        self,
        input_ids=None,
        attention_mask=None,
        token_type_ids=None,
        position_ids=None,
        head_mask=None,
        inputs_embeds=None,
        labels=None,
        output_attentions=None,
        output_hidden_states=None,
        return_tuple=None,
        **kwargs
    ):
        r"""
        labels (:obj:`torch.LongTensor` of shape :obj:`(batch_size, sequence_length)`, `optional`, defaults to :obj:`None`):
            Labels for computing the masked language modeling loss.
            Indices should be in ``[-100, 0, ..., config.vocab_size]`` (see ``input_ids`` docstring)
            Tokens with indices set to ``-100`` are ignored (masked), the loss is only computed for the tokens with labels
            in ``[0, ..., config.vocab_size]``
        kwargs (:obj:`Dict[str, any]`, optional, defaults to `{}`):
            Used to hide legacy arguments that have been deprecated.
        """
        if "masked_lm_labels" in kwargs:
            warnings.warn(
                "The `masked_lm_labels` argument is deprecated and will be removed in a future version, use `labels` instead.",
                DeprecationWarning,
            )
            labels = kwargs.pop("masked_lm_labels")
        assert kwargs == {}, f"Unexpected keyword arguments: {list(kwargs.keys())}."
        return_tuple = return_tuple if return_tuple is not None else self.config.use_return_tuple

        outputs = self.roberta(
            input_ids,
            attention_mask=attention_mask,
            token_type_ids=token_type_ids,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_tuple=return_tuple,
        )
        sequence_output = outputs[0]
        prediction_scores = self.lm_head(sequence_output)

        masked_lm_loss = None
        if labels is not None:
            loss_fct = CrossEntropyLoss()
            masked_lm_loss = loss_fct(prediction_scores.view(-1, self.config.vocab_size), labels.view(-1))

        if return_tuple:
            output = (prediction_scores,) + outputs[2:]
            return ((masked_lm_loss,) + output) if masked_lm_loss is not None else output

        return MaskedLMOutput(
            loss=masked_lm_loss,
            logits=prediction_scores,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
        )


class RobertaLMHead(nn.Module):
    """Roberta Head for masked language modeling."""

    def __init__(self, config):
        super().__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.layer_norm = BertLayerNorm(config.hidden_size, eps=config.layer_norm_eps)

        self.decoder = nn.Linear(config.hidden_size, config.vocab_size, bias=False)
        self.bias = nn.Parameter(torch.zeros(config.vocab_size))

        # Need a link between the two variables so that the bias is correctly resized with `resize_token_embeddings`
        self.decoder.bias = self.bias

    def forward(self, features, **kwargs):
        x = self.dense(features)
        x = gelu(x)
        x = self.layer_norm(x)

        # project back to size of vocabulary with bias
        x = self.decoder(x)

        return x


@add_start_docstrings(
    """RoBERTa Model transformer with a sequence classification/regression head on top (a linear layer
    on top of the pooled output) e.g. for GLUE tasks. """,
    ROBERTA_START_DOCSTRING,
)
class RobertaForSequenceClassification(BertPreTrainedModel):
    config_class = RobertaConfig
    base_model_prefix = "roberta"

    def __init__(self, config):
        super().__init__(config)
        self.num_labels = config.num_labels

        self.roberta = RobertaModel(config)
        self.classifier = RobertaClassificationHead(config)

        self.init_weights()

    @add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("(batch_size, sequence_length)"))
    @add_code_sample_docstrings(
        tokenizer_class=_TOKENIZER_FOR_DOC,
        checkpoint="roberta-base",
        output_type=SequenceClassifierOutput,
        config_class=_CONFIG_FOR_DOC,
    )
    def forward(
        self,
        input_ids=None,
        attention_mask=None,
        token_type_ids=None,
        position_ids=None,
        head_mask=None,
        inputs_embeds=None,
        labels=None,
        output_attentions=None,
        output_hidden_states=None,
        return_tuple=None,
    ):
        r"""
        labels (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`, defaults to :obj:`None`):
            Labels for computing the sequence classification/regression loss.
            Indices should be in :obj:`[0, ..., config.num_labels - 1]`.
            If :obj:`config.num_labels == 1` a regression loss is computed (Mean-Square loss),
            If :obj:`config.num_labels > 1` a classification loss is computed (Cross-Entropy).
        """
        return_tuple = return_tuple if return_tuple is not None else self.config.use_return_tuple

        outputs = self.roberta(
            input_ids,
            attention_mask=attention_mask,
            token_type_ids=token_type_ids,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_tuple=return_tuple,
        )
        sequence_output = outputs[0]
        logits = self.classifier(sequence_output)

        loss = None
        if labels is not None:
            if self.num_labels == 1:
                #  We are doing regression
                loss_fct = MSELoss()
                loss = loss_fct(logits.view(-1), labels.view(-1))
            else:
                loss_fct = CrossEntropyLoss()
                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))

        if return_tuple:
            output = (logits,) + outputs[2:]
            return ((loss,) + output) if loss is not None else output

        return SequenceClassifierOutput(
            loss=loss, logits=logits, hidden_states=outputs.hidden_states, attentions=outputs.attentions,
        )


@add_start_docstrings(
    """Roberta 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. """,
    ROBERTA_START_DOCSTRING,
)
class RobertaForMultipleChoice(BertPreTrainedModel):
    config_class = RobertaConfig
    base_model_prefix = "roberta"

    def __init__(self, config):
        super().__init__(config)

        self.roberta = RobertaModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, 1)

        self.init_weights()

    @add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("(batch_size, num_choices, sequence_length)"))
    @add_code_sample_docstrings(
        tokenizer_class=_TOKENIZER_FOR_DOC,
        checkpoint="roberta-base",
        output_type=MultipleChoiceModelOutput,
        config_class=_CONFIG_FOR_DOC,
    )
    def forward(
        self,
        input_ids=None,
        token_type_ids=None,
        attention_mask=None,
        labels=None,
        position_ids=None,
        head_mask=None,
        inputs_embeds=None,
        output_attentions=None,
        output_hidden_states=None,
        return_tuple=None,
    ):
        r"""
        labels (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`, defaults to :obj:`None`):
            Labels for computing the multiple choice classification loss.
            Indices should be in ``[0, ..., num_choices]`` where `num_choices` is the size of the second dimension
            of the input tensors. (see `input_ids` above)
        """
        return_tuple = return_tuple if return_tuple is not None else self.config.use_return_tuple
        num_choices = input_ids.shape[1] if input_ids is not None else inputs_embeds.shape[1]

        flat_input_ids = input_ids.view(-1, input_ids.size(-1)) if input_ids is not None else None
        flat_position_ids = position_ids.view(-1, position_ids.size(-1)) if position_ids is not None else None
        flat_token_type_ids = token_type_ids.view(-1, token_type_ids.size(-1)) if token_type_ids is not None else None
        flat_attention_mask = attention_mask.view(-1, attention_mask.size(-1)) if attention_mask is not None else None
        flat_inputs_embeds = (
            inputs_embeds.view(-1, inputs_embeds.size(-2), inputs_embeds.size(-1))
            if inputs_embeds is not None
            else None
        )

        outputs = self.roberta(
            flat_input_ids,
            position_ids=flat_position_ids,
            token_type_ids=flat_token_type_ids,
            attention_mask=flat_attention_mask,
            head_mask=head_mask,
            inputs_embeds=flat_inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_tuple=return_tuple,
        )
        pooled_output = outputs[1]

        pooled_output = self.dropout(pooled_output)
        logits = self.classifier(pooled_output)
        reshaped_logits = logits.view(-1, num_choices)

        loss = None
        if labels is not None:
            loss_fct = CrossEntropyLoss()
            loss = loss_fct(reshaped_logits, labels)

        if return_tuple:
            output = (reshaped_logits,) + outputs[2:]
            return ((loss,) + output) if loss is not None else output

        return MultipleChoiceModelOutput(
            loss=loss, logits=reshaped_logits, hidden_states=outputs.hidden_states, attentions=outputs.attentions,
        )


@add_start_docstrings(
    """Roberta 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. """,
    ROBERTA_START_DOCSTRING,
)
class RobertaForTokenClassification(BertPreTrainedModel):
    config_class = RobertaConfig
    base_model_prefix = "roberta"

    def __init__(self, config):
        super().__init__(config)
        self.num_labels = config.num_labels

        self.roberta = RobertaModel(config)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.classifier = nn.Linear(config.hidden_size, config.num_labels)

        self.init_weights()

    @add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("(batch_size, sequence_length)"))
    @add_code_sample_docstrings(
        tokenizer_class=_TOKENIZER_FOR_DOC,
        checkpoint="roberta-base",
        output_type=TokenClassifierOutput,
        config_class=_CONFIG_FOR_DOC,
    )
    def forward(
        self,
        input_ids=None,
        attention_mask=None,
        token_type_ids=None,
        position_ids=None,
        head_mask=None,
        inputs_embeds=None,
        labels=None,
        output_attentions=None,
        output_hidden_states=None,
        return_tuple=None,
    ):
        r"""
        labels (:obj:`torch.LongTensor` of shape :obj:`(batch_size, sequence_length)`, `optional`, defaults to :obj:`None`):
            Labels for computing the token classification loss.
            Indices should be in ``[0, ..., config.num_labels - 1]``.
        """
        return_tuple = return_tuple if return_tuple is not None else self.config.use_return_tuple

        outputs = self.roberta(
            input_ids,
            attention_mask=attention_mask,
            token_type_ids=token_type_ids,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_tuple=return_tuple,
        )

        sequence_output = outputs[0]

        sequence_output = self.dropout(sequence_output)
        logits = self.classifier(sequence_output)

        loss = None
        if labels is not None:
            loss_fct = CrossEntropyLoss()
            # Only keep active parts of the loss
            if attention_mask is not None:
                active_loss = attention_mask.view(-1) == 1
                active_logits = logits.view(-1, self.num_labels)
                active_labels = torch.where(
                    active_loss, labels.view(-1), torch.tensor(loss_fct.ignore_index).type_as(labels)
                )
                loss = loss_fct(active_logits, active_labels)
            else:
                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))

        if return_tuple:
            output = (logits,) + outputs[2:]
            return ((loss,) + output) if loss is not None else output

        return TokenClassifierOutput(
            loss=loss, logits=logits, hidden_states=outputs.hidden_states, attentions=outputs.attentions,
        )


class RobertaClassificationHead(nn.Module):
    """Head for sentence-level classification tasks."""

    def __init__(self, config):
        super().__init__()
        self.dense = nn.Linear(config.hidden_size, config.hidden_size)
        self.dropout = nn.Dropout(config.hidden_dropout_prob)
        self.out_proj = nn.Linear(config.hidden_size, config.num_labels)

    def forward(self, features, **kwargs):
        x = features[:, 0, :]  # take <s> token (equiv. to [CLS])
        x = self.dropout(x)
        x = self.dense(x)
        x = torch.tanh(x)
        x = self.dropout(x)
        x = self.out_proj(x)
        return x


@add_start_docstrings(
    """Roberta 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`). """,
    ROBERTA_START_DOCSTRING,
)
class RobertaForQuestionAnswering(BertPreTrainedModel):
    config_class = RobertaConfig
    base_model_prefix = "roberta"

    def __init__(self, config):
        super().__init__(config)
        self.num_labels = config.num_labels

        self.roberta = RobertaModel(config)
        self.qa_outputs = nn.Linear(config.hidden_size, config.num_labels)

        self.init_weights()

    @add_start_docstrings_to_callable(ROBERTA_INPUTS_DOCSTRING.format("(batch_size, sequence_length)"))
    @add_code_sample_docstrings(
        tokenizer_class=_TOKENIZER_FOR_DOC,
        checkpoint="roberta-base",
        output_type=QuestionAnsweringModelOutput,
        config_class=_CONFIG_FOR_DOC,
    )
    def forward(
        self,
        input_ids=None,
        attention_mask=None,
        token_type_ids=None,
        position_ids=None,
        head_mask=None,
        inputs_embeds=None,
        start_positions=None,
        end_positions=None,
        output_attentions=None,
        output_hidden_states=None,
        return_tuple=None,
    ):
        r"""
        start_positions (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`, defaults to :obj:`None`):
            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 (:obj:`torch.LongTensor` of shape :obj:`(batch_size,)`, `optional`, defaults to :obj:`None`):
            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.
        """
        return_tuple = return_tuple if return_tuple is not None else self.config.use_return_tuple

        outputs = self.roberta(
            input_ids,
            attention_mask=attention_mask,
            token_type_ids=token_type_ids,
            position_ids=position_ids,
            head_mask=head_mask,
            inputs_embeds=inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_tuple=return_tuple,
        )

        sequence_output = outputs[0]

        logits = self.qa_outputs(sequence_output)
        start_logits, end_logits = logits.split(1, dim=-1)
        start_logits = start_logits.squeeze(-1)
        end_logits = end_logits.squeeze(-1)

        total_loss = None
        if start_positions is not None and end_positions is not None:
            # If we are on multi-GPU, split add a dimension
            if len(start_positions.size()) > 1:
                start_positions = start_positions.squeeze(-1)
            if len(end_positions.size()) > 1:
                end_positions = end_positions.squeeze(-1)
            # sometimes the start/end positions are outside our model inputs, we ignore these terms
            ignored_index = start_logits.size(1)
            start_positions.clamp_(0, ignored_index)
            end_positions.clamp_(0, ignored_index)

            loss_fct = CrossEntropyLoss(ignore_index=ignored_index)
            start_loss = loss_fct(start_logits, start_positions)
            end_loss = loss_fct(end_logits, end_positions)
            total_loss = (start_loss + end_loss) / 2

        if return_tuple:
            output = (start_logits, end_logits) + outputs[2:]
            return ((total_loss,) + output) if total_loss is not None else output

        return QuestionAnsweringModelOutput(
            loss=total_loss,
            start_logits=start_logits,
            end_logits=end_logits,
            hidden_states=outputs.hidden_states,
            attentions=outputs.attentions,
        )


def create_position_ids_from_input_ids(input_ids, padding_idx):
    """ Replace non-padding symbols with their position numbers. Position numbers begin at
    padding_idx+1. Padding symbols are ignored. This is modified from fairseq's
    `utils.make_positions`.

    :param torch.Tensor x:
    :return torch.Tensor:
    """
    # The series of casts and type-conversions here are carefully balanced to both work with ONNX export and XLA.
    mask = input_ids.ne(padding_idx).int()
    incremental_indices = torch.cumsum(mask, dim=1).type_as(mask) * mask
    return incremental_indices.long() + padding_idx