batch_stats.update(stats)
return None, batch_stats
- def _stats(self, loss, scores, target):
+ def _stats(self, loss, scores, labels):
"""
Args:
loss (:obj:`FloatTensor`): the loss computed by the loss criterion.
scores (:obj:`FloatTensor`): a score for each possible output
- target (:obj:`FloatTensor`): true targets
+ labels (:obj:`FloatTensor`): true targets
Returns:
:obj:`onmt.utils.Statistics` : statistics for this batch.
"""
pred = scores.max(1)[1]
- non_padding = target.ne(self.padding_idx)
- num_correct = pred.eq(target).masked_select(non_padding).sum().item()
+ non_padding = labels.ne(self.padding_idx)
+ num_correct = pred.eq(labels).masked_select(non_padding).sum().item()
num_non_padding = non_padding.sum().item()
return onmt.utils.Statistics(loss.item(), num_non_padding, num_correct)
@@ -399,14 +399,14 @@ Source code for onmt.utils.loss
self.confidence = 1.0 - label_smoothing
- def forward(self, output, target):
+ def forward(self, output, labels):
"""
output (FloatTensor): batch_size x n_classes
- target (LongTensor): batch_size
+ labels (LongTensor): batch_size
"""
- model_prob = self.one_hot.repeat(target.size(0), 1)
- model_prob.scatter_(1, target.unsqueeze(1), self.confidence)
- model_prob.masked_fill_((target == self.ignore_index).unsqueeze(1), 0)
+ model_prob = self.one_hot.repeat(labels.size(0), 1)
+ model_prob.scatter_(1, labels.unsqueeze(1), self.confidence)
+ model_prob.masked_fill_((labels == self.ignore_index).unsqueeze(1), 0)
return F.kl_div(output, model_prob, reduction='sum')