Model Card for mdeberta-v3-base-triplet-critic-xnli

This is the Triplit Critic model presented in the ACL 2023 paper RED^{FM}: a Filtered and Multilingual Relation Extraction Dataset. If you use the model, please reference this work in your paper:

@inproceedings{huguet-cabot-et-al-2023-redfm-dataset,
    title = "RED$^{\rm FM}$: a Filtered and Multilingual Relation Extraction Dataset",
    author = "Huguet Cabot, Pere-Llu{\'\i}s  and Tedeschi, Simone and Ngonga Ngomo, Axel-Cyrille and
      Navigli, Roberto",
    booktitle = "Proc. of the 61st Annual Meeting of the Association for Computational Linguistics: ACL 2023",
    month = jul,
    year = "2023",
    address = "Toronto, Canada",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/2306.09802",
}

The Triplit Critic is based on mdeberta-v3-base and it was trained as a multitask system to filter triplets (triplet filtering data used can be found in Triplet-Critic-Data) as well as on the XNLI dataset. The model weights contain the two classification heads, however loading it using the huggingface library will only load those for Triplet filtering (ie. a binary classification head), if one wants to use it for XNLI it needs a custom script. While it is defined and trained as a classification system, we use the positive score (ie. Label_1) as the confidence score for a triplet. For SRED^FM the confidence score thresshold was set at 0.75.

To load the multitask model:

from transformers import DebertaV2PreTrainedModel, DebertaV2Model
from torch import nn
from transformers.models.deberta_v2.modeling_deberta_v2 import *
from transformers.file_utils import ModelOutput

@dataclass
class TXNLIClassifierOutput(ModelOutput):
    """
    Base class for outputs of sentence classification models.

    Args:
        loss (:obj:`torch.FloatTensor` of shape :obj:`(1,)`, `optional`, returned when :obj:`labels` is provided):
            Classification (or regression if config.num_labels==1) loss.
        logits (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, config.num_labels)`):
            Classification (or regression if config.num_labels==1) scores (before SoftMax).
        logits_xnli (:obj:`torch.FloatTensor` of shape :obj:`(batch_size, config.num_labels)`):
            Classification (or regression if config.num_labels==1) scores (before SoftMax).
        hidden_states (:obj:`tuple(torch.FloatTensor)`, `optional`, returned when ``output_hidden_states=True`` is passed or when ``config.output_hidden_states=True``):
            Tuple of :obj:`torch.FloatTensor` (one for the output of the embeddings + one for the output of each layer)
            of shape :obj:`(batch_size, sequence_length, hidden_size)`.

            Hidden-states of the model at the output of each layer plus the initial embedding outputs.
        attentions (:obj:`tuple(torch.FloatTensor)`, `optional`, returned when ``output_attentions=True`` is passed or when ``config.output_attentions=True``):
            Tuple of :obj:`torch.FloatTensor` (one for each layer) of shape :obj:`(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.
    """

    loss: Optional[torch.FloatTensor] = None
    logits: torch.FloatTensor = None
    logits_xnli: torch.FloatTensor = None
    hidden_states: Optional[Tuple[torch.FloatTensor]] = None
    attentions: Optional[Tuple[torch.FloatTensor]] = None

class DebertaV2ForTripletClassification(DebertaV2PreTrainedModel):
    def __init__(self, config):
        super().__init__(config)

        num_labels = getattr(config, "num_labels", 2)
        self.num_labels = num_labels

        self.deberta = DebertaV2Model(config)
        self.pooler = ContextPooler(config)
        output_dim = self.pooler.output_dim

        self.classifier = nn.Linear(output_dim, num_labels)
        drop_out = getattr(config, "cls_dropout", None)
        drop_out = self.config.hidden_dropout_prob if drop_out is None else drop_out
        self.dropout = StableDropout(drop_out)
        self.classifier_xnli = nn.Linear(output_dim, 3)

        # Initialize weights and apply final processing
        self.post_init()

    def get_input_embeddings(self):
        return self.deberta.get_input_embeddings()

    def set_input_embeddings(self, new_embeddings):
        self.deberta.set_input_embeddings(new_embeddings)

    @add_start_docstrings_to_model_forward(DEBERTA_INPUTS_DOCSTRING.format("batch_size, sequence_length"))
    def forward(
        self,
        input_ids=None,
        attention_mask=None,
        token_type_ids=None,
        position_ids=None,
        inputs_embeds=None,
        labels=None,
        output_attentions=None,
        output_hidden_states=None,
        return_dict=None,
    ):
        r"""
        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]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss),
            If `config.num_labels > 1` a classification loss is computed (Cross-Entropy).
        """
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        outputs = self.deberta(
            input_ids,
            token_type_ids=token_type_ids,
            attention_mask=attention_mask,
            position_ids=position_ids,
            inputs_embeds=inputs_embeds,
            output_attentions=output_attentions,
            output_hidden_states=output_hidden_states,
            return_dict=return_dict,
        )

        encoder_layer = outputs[0]
        pooled_output = self.pooler(encoder_layer)
        pooled_output = self.dropout(pooled_output)
        logits = self.classifier(pooled_output)
        logits_xnli = self.classifier_xnli(pooled_output)

        loss = None
        if labels is not None:
            if labels.dtype != torch.bool:
                loss_fct = CrossEntropyLoss()
                loss = loss_fct(logits.view(-1, self.num_labels), labels.view(-1))
            else:
                loss_fct = BCEWithLogitsLoss()
                loss = loss_fct(logits_xnli.view(-1, 3), labels.view(-1).long())
        if not return_dict:
            output = (logits,) + outputs[1:]
            return ((loss,) + output) if loss is not None else output

        return TXNLIClassifierOutput(
            loss=loss, logits=logits, logits_xnli=logits_xnli, hidden_states=outputs.hidden_states, attentions=outputs.attentions
        )

License

This model is licensed under the CC BY-SA 4.0 license. The text of the license can be found here.