TripletLoss.py

from enum import Enum
from typing import Dict, Iterable

import torch.nn.functional as F
from torch import Tensor, nn

from sentence_transformers.SentenceTransformer import SentenceTransformer


class TripletDistanceMetric(Enum):
    """The metric for the triplet loss"""

    COSINE = lambda x, y: 1 - F.cosine_similarity(x, y)
    EUCLIDEAN = lambda x, y: F.pairwise_distance(x, y, p=2)
    MANHATTAN = lambda x, y: F.pairwise_distance(x, y, p=1)


class TripletLoss(nn.Module):
    def __init__(
        self, model: SentenceTransformer, distance_metric=TripletDistanceMetric.EUCLIDEAN, triplet_margin: float = 5
    ):
        """
        This class implements triplet loss. Given a triplet of (anchor, positive, negative),
        the loss minimizes the distance between anchor and positive while it maximizes the distance
        between anchor and negative. It compute the following loss function:

        ``loss = max(||anchor - positive|| - ||anchor - negative|| + margin, 0)``.

        Margin is an important hyperparameter and needs to be tuned respectively.

        Args:
            model: SentenceTransformerModel
            distance_metric: Function to compute distance between two
                embeddings. The class TripletDistanceMetric contains
                common distance metrices that can be used.
            triplet_margin: The negative should be at least this much
                further away from the anchor than the positive.

        References:
            - For further details, see: https://en.wikipedia.org/wiki/Triplet_loss

        Requirements:
            1. (anchor, positive, negative) triplets

        Inputs:
            +---------------------------------------+--------+
            | Texts                                 | Labels |
            +=======================================+========+
            | (anchor, positive, negative) triplets | none   |
            +---------------------------------------+--------+

        Example:
            ::

                from sentence_transformers import SentenceTransformer, SentenceTransformerTrainer, losses
                from datasets import Dataset

                model = SentenceTransformer("microsoft/mpnet-base")
                train_dataset = Dataset.from_dict({
                    "anchor": ["It's nice weather outside today.", "He drove to work."],
                    "positive": ["It's so sunny.", "He took the car to the office."],
                    "negative": ["It's quite rainy, sadly.", "She walked to the store."],
                })
                loss = losses.TripletLoss(model=model)

                trainer = SentenceTransformerTrainer(
                    model=model,
                    train_dataset=train_dataset,
                    loss=loss,
                )
                trainer.train()
        """
        super(TripletLoss, self).__init__()
        self.model = model
        self.distance_metric = distance_metric
        self.triplet_margin = triplet_margin

    def forward(self, sentence_features: Iterable[Dict[str, Tensor]], labels: Tensor):
        reps = [self.model(sentence_feature)["sentence_embedding"] for sentence_feature in sentence_features]

        rep_anchor, rep_pos, rep_neg = reps
        distance_pos = self.distance_metric(rep_anchor, rep_pos)
        distance_neg = self.distance_metric(rep_anchor, rep_neg)

        losses = F.relu(distance_pos - distance_neg + self.triplet_margin)
        return losses.mean()

    def get_config_dict(self):
        distance_metric_name = self.distance_metric.__name__
        for name, value in vars(TripletDistanceMetric).items():
            if value == self.distance_metric:
                distance_metric_name = "TripletDistanceMetric.{}".format(name)
                break

        return {"distance_metric": distance_metric_name, "triplet_margin": self.triplet_margin}

    @property
    def citation(self) -> str:
        return """
@misc{hermans2017defense,
    title={In Defense of the Triplet Loss for Person Re-Identification}, 
    author={Alexander Hermans and Lucas Beyer and Bastian Leibe},
    year={2017},
    eprint={1703.07737},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}
"""