# Copyright 2020 - 2021 MONAI Consortium
# 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.
from typing import Union
import torch
from torch.nn import functional as F
from torch.nn.modules.loss import _Loss
from monai.utils import LossReduction
[docs]class ContrastiveLoss(_Loss):
"""
Compute the Contrastive loss defined in:
Chen, Ting, et al. "A simple framework for contrastive learning of visual representations." International
conference on machine learning. PMLR, 2020. (http://proceedings.mlr.press/v119/chen20j.html)
Adapted from:
https://github.com/Sara-Ahmed/SiT/blob/1aacd6adcd39b71efc903d16b4e9095b97dda76f/losses.py#L5
"""
[docs] def __init__(
self, temperature: float = 0.5, batch_size: int = 1, reduction: Union[LossReduction, str] = LossReduction.SUM
) -> None:
"""
Args:
temperature: Can be scaled between 0 and 1 for learning from negative samples, ideally set to 0.5.
Raises:
AssertionError: When an input of dimension length > 2 is passed
AssertionError: When input and target are of different shapes
"""
super().__init__(reduction=LossReduction(reduction).value)
self.batch_size = batch_size
self.temperature = temperature
[docs] def forward(self, input: torch.Tensor, target: torch.Tensor) -> torch.Tensor:
"""
Args:
input: the shape should be B[F].
target: the shape should be B[F].
Raises:
ValueError: When ``self.reduction`` is not one of ["sum", "none"].
"""
if len(target.shape) > 2 or len(input.shape) > 2:
raise AssertionError(
f"Either target or input has dimensions greater than 2 where target "
f"shape is ({target.shape}) and input shape is ({input.shape})"
)
if target.shape != input.shape:
raise AssertionError(f"ground truth has differing shape ({target.shape}) from input ({input.shape})")
temperature_tensor = torch.tensor(self.temperature).to(input.device)
norm_i = F.normalize(input, dim=1)
norm_j = F.normalize(target, dim=1)
negatives_mask = ~torch.eye(self.batch_size * 2, self.batch_size * 2, dtype=torch.bool)
negatives_mask = torch.tensor(negatives_mask, dtype=torch.float)
negatives_mask = torch.clone(torch.as_tensor(negatives_mask)).to(input.device)
repr = torch.cat([norm_i, norm_j], dim=0)
sim_matrix = F.cosine_similarity(repr.unsqueeze(1), repr.unsqueeze(0), dim=2)
sim_ij = torch.diag(sim_matrix, self.batch_size)
sim_ji = torch.diag(sim_matrix, -self.batch_size)
positives = torch.cat([sim_ij, sim_ji], dim=0)
nominator = torch.exp(positives / temperature_tensor)
denominator = negatives_mask * torch.exp(sim_matrix / temperature_tensor)
loss_partial = -torch.log(nominator / torch.sum(denominator, dim=1))
if self.reduction == LossReduction.SUM.value:
return torch.sum(loss_partial) / (2 * self.batch_size)
raise ValueError(f"Unsupported reduction: {self.reduction}, " f'available options are ["mean", "sum", "none"].')