Lines Matching refs:losses
15 """Tests for losses."""
34 from tensorflow.python.ops.losses import losses
35 from tensorflow.python.ops.losses import util
50 losses.absolute_difference(
54 loss = losses.absolute_difference(self._predictions, self._predictions)
59 loss = losses.absolute_difference(self._labels, self._predictions)
65 loss = losses.absolute_difference(self._labels, self._predictions, weights)
71 loss = losses.absolute_difference(self._labels, self._predictions,
78 loss = losses.absolute_difference(self._labels, self._predictions, weights)
84 loss = losses.absolute_difference(self._labels, self._predictions, weights)
90 loss = losses.absolute_difference(self._labels, self._predictions, weights)
96 loss = losses.absolute_difference(self._labels, self._predictions, weights)
102 loss = losses.absolute_difference(self._labels, self._predictions, weights)
115 losses.softmax_cross_entropy(labels, logits, weights=None)
122 loss = losses.softmax_cross_entropy(labels, logits)
132 loss = losses.softmax_cross_entropy(labels, logits)
142 loss = losses.softmax_cross_entropy(labels, logits, weights)
151 loss = losses.softmax_cross_entropy(labels, logits,
161 loss = losses.softmax_cross_entropy(labels, logits, weights)
170 loss = losses.softmax_cross_entropy(labels, logits, weights)
179 loss = losses.softmax_cross_entropy(labels, logits, weights)
191 losses.softmax_cross_entropy(labels, logits, weights=weights).eval()
208 loss = losses.softmax_cross_entropy(
223 losses.sparse_softmax_cross_entropy(labels, logits, weights=None)
230 loss = losses.sparse_softmax_cross_entropy(labels, logits)
239 loss = losses.sparse_softmax_cross_entropy(labels, logits)
248 loss = losses.sparse_softmax_cross_entropy(labels, logits)
258 loss = losses.sparse_softmax_cross_entropy(labels, logits)
268 loss = losses.sparse_softmax_cross_entropy(labels, logits)
278 loss = losses.sparse_softmax_cross_entropy(labels, logits)
288 loss = losses.sparse_softmax_cross_entropy(labels, logits, weights)
297 loss = losses.sparse_softmax_cross_entropy(labels, logits,
307 loss = losses.sparse_softmax_cross_entropy(
318 loss = losses.sparse_softmax_cross_entropy(labels, logits, weights)
328 loss = losses.sparse_softmax_cross_entropy(labels, logits, weights)
343 loss = losses.sparse_softmax_cross_entropy(labels, logits, weights)
360 loss = losses.sparse_softmax_cross_entropy(labels, logits, weights)
369 loss = losses.sparse_softmax_cross_entropy(labels, logits, weights)
378 loss = losses.sparse_softmax_cross_entropy(labels, logits, weights)
387 loss = losses.sparse_softmax_cross_entropy(labels, logits, weights)
399 losses.sparse_softmax_cross_entropy(
412 losses.sparse_softmax_cross_entropy(
425 losses.sparse_softmax_cross_entropy(
439 losses.sparse_softmax_cross_entropy(
453 losses.sparse_softmax_cross_entropy(
465 loss = losses.sigmoid_cross_entropy(labels, logits)
475 loss = losses.sigmoid_cross_entropy(labels, logits, weights)
491 loss = losses.sigmoid_cross_entropy(labels, logits, weights)
508 loss = losses.sigmoid_cross_entropy(labels, logits)
520 loss = losses.sigmoid_cross_entropy(labels, logits, weights)
530 loss = losses.sigmoid_cross_entropy(labels, logits)
546 loss = losses.sigmoid_cross_entropy(labels, logits)
558 loss = losses.sigmoid_cross_entropy(
559 labels, logits, reduction=losses.Reduction.NONE)
585 loss = losses.sigmoid_cross_entropy(
597 sigmoid_loss = losses.sigmoid_cross_entropy(
604 softmax_loss = losses.softmax_cross_entropy(
630 losses.log_loss(self._labels, self._labels, weights=None)
633 loss = losses.log_loss(self._labels, self._labels)
640 loss = losses.log_loss(self._labels, tf_predictions)
646 loss = losses.log_loss(self._labels, self._predictions)
653 loss = losses.log_loss(self._labels, self._predictions, weights)
660 loss = losses.log_loss(self._labels, self._predictions,
670 loss = losses.log_loss(self._labels, tf_predictions,
680 loss = losses.log_loss(self._labels, tf_predictions,
692 loss = losses.log_loss(self._labels, self._predictions, weights)
701 loss = losses.log_loss(self._labels, self._predictions, weights)
710 loss = losses.log_loss(self._labels, self._predictions, weights)
718 losses.log_loss(self._labels, self._predictions, weights)
724 loss = losses.log_loss(
737 loss = losses.log_loss(
751 loss = losses.log_loss(
765 loss = losses.log_loss(self._labels, tf_predictions, tf_weights)
773 loss = losses.log_loss(self._labels, self._predictions, tf_weights)
785 _ = losses.hinge_loss(labels, logits).eval()
791 loss = losses.hinge_loss(labels, logits)
798 loss = losses.hinge_loss(labels, logits)
807 loss = losses.hinge_loss(labels, logits)
820 _ = losses.huber_loss(labels, predictions).eval()
826 loss = losses.huber_loss(labels, predictions)
834 loss = losses.huber_loss(labels, predictions)
844 loss = losses.huber_loss(labels, predictions)
856 loss = losses.huber_loss(labels, predictions, delta=delta)
865 loss = losses.huber_loss(labels, predictions, delta=delta)
880 losses.mean_squared_error(
887 losses.mean_squared_error(predictions=constant_op.constant(0),
891 loss = losses.mean_squared_error(self._predictions, self._predictions)
896 loss = losses.mean_squared_error(self._labels, self._predictions)
902 loss = losses.mean_squared_error(self._labels, self._predictions, weights)
908 loss = losses.mean_squared_error(self._labels, self._predictions,
915 loss = losses.mean_squared_error(self._labels, self._predictions, weights)
921 loss = losses.mean_squared_error(self._labels, self._predictions, weights)
927 loss = losses.mean_squared_error(self._labels, self._predictions, weights)
933 loss = losses.mean_squared_error(self._labels, self._predictions, weights)
939 loss = losses.mean_squared_error(self._labels, self._predictions, weights)
968 losses.mean_pairwise_squared_error(
976 static_inputs_op = losses.mean_pairwise_squared_error(
986 dynamic_inputs_op = losses.mean_pairwise_squared_error(
1020 loss = losses.mean_pairwise_squared_error(predictions, predictions, 0)
1041 loss = losses.mean_pairwise_squared_error(
1083 losses.mean_pairwise_squared_error(
1090 dynamic_inputs_op = losses.mean_pairwise_squared_error(
1153 losses.mean_pairwise_squared_error(
1156 loss1 = losses.mean_pairwise_squared_error(
1159 loss0_1 = losses.mean_pairwise_squared_error(
1190 losses.cosine_distance(
1197 loss = losses.cosine_distance(
1205 loss = losses.cosine_distance(
1225 loss = losses.cosine_distance(tf_labels, tf_preds, dim=2)
1231 loss = losses.cosine_distance(
1240 loss = losses.cosine_distance(
1252 loss = losses.cosine_distance(
1263 loss = losses.cosine_distance(
1272 loss = losses.cosine_distance(
1287 losses.absolute_difference(logits, labels, loss_collection=None)
1288 losses.log_loss(logits, labels, loss_collection=None)
1289 losses.mean_squared_error(logits, labels, loss_collection=None)
1290 losses.sigmoid_cross_entropy(logits, labels, loss_collection=None)
1291 losses.softmax_cross_entropy(logits, labels, loss_collection=None)
1311 for reduction in losses.Reduction.all():
1316 losses.compute_weighted_loss(raw_losses, reduction=reduction),
1317 losses.compute_weighted_loss(
1319 losses.compute_weighted_loss(
1321 losses.compute_weighted_loss(
1323 losses.compute_weighted_loss(
1325 losses.compute_weighted_loss(
1327 losses.compute_weighted_loss(
1329 losses.compute_weighted_loss(
1331 losses.compute_weighted_loss(
1337 if reduction == losses.Reduction.NONE:
1339 elif reduction == losses.Reduction.SUM:
1349 for reduction in losses.Reduction.all():
1355 losses.compute_weighted_loss(raw_losses, reduction=reduction),
1356 losses.compute_weighted_loss(
1358 losses.compute_weighted_loss(
1364 if reduction == losses.Reduction.NONE:
1367 elif reduction == losses.Reduction.SUM:
1380 weighted_loss = losses.compute_weighted_loss(
1394 losses.compute_weighted_loss(self._raw_losses, weights=weights)
1398 weighted_loss = losses.compute_weighted_loss(
1417 losses.compute_weighted_loss(raw_losses, weights=weights)
1421 weighted_loss = losses.compute_weighted_loss(
1447 for reduction in losses.Reduction.all():
1450 weighted_loss = losses.compute_weighted_loss(
1456 if reduction == losses.Reduction.NONE:
1458 elif reduction == losses.Reduction.SUM:
1462 if reduction == losses.Reduction.MEAN:
1466 elif (reduction == losses.Reduction.SUM_OVER_NONZERO_WEIGHTS or
1467 reduction == losses.Reduction.SUM_BY_NONZERO_WEIGHTS):
1471 elif reduction == losses.Reduction.SUM_OVER_BATCH_SIZE:
