Function binary_cross_entropy_with_logits

loss.py:161–187

(input, target, weight=None, size_average=None,
                                     reduce=False, reduction='elementwise_mean', pos_weight=None,mask=None)

    return loss

def binary_cross_entropy_with_logits(input, target, weight=None, size_average=None,
                                     reduce=False, reduction='elementwise_mean', pos_weight=None,mask=None):

    if not (target.size() == input.size()):
        raise ValueError("Target size ({}) must be the same as input size ({})".format(target.size(), input.size()))

    max_val = (-input).clamp(min=0)

    if pos_weight is None:
        ce_loss = input - input * target + max_val + ((-max_val).exp() + (-input - max_val).exp()).log()
    else:
        log_weight = 1 + (pos_weight - 1) * target
        ce_loss = input - input * target + log_weight * (max_val + ((-max_val).exp() + (-input - max_val).exp()).log())

            
    if weight is not None:
        ce_loss = ce_loss * weight
    if mask is not None:
        
        ce_loss = ce_loss[mask.unsqueeze(1).repeat(1,ce_loss.shape[1],1,1)]

    if reduction == False:
        return ce_loss
    elif reduction == 'elementwise_mean':
        return ce_loss.mean()
    else:
        return ce_loss.sum()
def classification_loss_1(prob_volume, depth_values, interval, depth_gt, mask, weight):
    depth_gt_volume = depth_gt.unsqueeze(1).expand_as(depth_values)  # (b, d, h, w)