ce69ccc9be6560ca1ccf82319a0af2bebde1ceda,onmt/utils/loss.py,NMTLossCompute,__init__,#NMTLossCompute#Any#Any#Any#Any#,180

Before Change

            one_hot = torch.randn(1, len(tgt_vocab))
            one_hot.fill_(label_smoothing / (len(tgt_vocab) - 2))
            one_hot[0][self.padding_idx] = 0
            self.register_buffer("one_hot", one_hot)
        else:
            weight = torch.ones(len(tgt_vocab))
            weight[self.padding_idx] = 0
            self.criterion = nn.NLLLoss(weight, size_average=False)

After Change

    def __init__(self, generator, tgt_vocab, normalization="sents",
                 label_smoothing=0.0):
        super(NMTLossCompute, self).__init__(generator, tgt_vocab)
        self.sparse = not isinstance(generator[1], nn.LogSoftmax)
        if label_smoothing > 0:
            self.criterion = LabelSmoothingLoss(
                label_smoothing, len(tgt_vocab), ignore_index=self.padding_idx

Instances