1:BPR
根据预测的pos_sorce和neg_sorce计算损失。
BPRLoss()
class BPRLoss(nn.Module):
def __init__(self, gamma=1e-10):
super(BPRLoss, self).__init__()
self.gamma = gamma
def forward(self, pos_score, neg_score):
loss = -torch.log(self.gamma + torch.sigmoid(pos_score - neg_score)).mean()
return loss
2:XDeepFM
根据lable=tensor([0.]),和intput = tensor([0.1999])计算损失。
self.loss = nn.BCELoss()
class BCELoss(_WeightedLoss):
return F.binary_cross_entropy(input, target, weight=self.weight, reduction=self.reduction)
根据参数计算L2损失
l2_reg = self.reg_loss(self.mlp_layers.named_parameters())
def