| /external/tensorflow/tensorflow/contrib/kfac/python/ops/ |
| curvature_matrix_vector_products.py | 35 loss function with respect to a provided list of tensors ("wrt_tensors").
36 For example, the Fisher associated with a log-prob loss w.r.t. the
62 losses: A list of LossFunction instances whose sum defines the total loss.
68 self._inputs_to_losses = list(loss.inputs for loss in losses)
74 return math_ops.add_n(tuple(loss.evaluate() for loss in self._losses))
97 """Multiply loss_vecs by Fisher of total loss."""
99 loss.multiply_fisher(loss_vec)
100 for loss, loss_vec in zip(self._losses, loss_vecs)
[all...] |
| /external/tensorflow/tensorflow/contrib/losses/python/losses/ |
| loss_ops_test.py | 54 loss = loss_ops.absolute_difference(self._predictions, self._predictions)
56 self.assertAlmostEqual(0.0, loss.eval(), 3)
59 loss = loss_ops.absolute_difference(self._predictions, self._labels)
61 self.assertAlmostEqual(5.5, loss.eval(), 3)
65 loss = loss_ops.absolute_difference(self._predictions, self._labels,
68 self.assertAlmostEqual(5.5 * weights, loss.eval(), 3)
72 loss = loss_ops.absolute_difference(self._predictions, self._labels,
75 self.assertAlmostEqual(5.5 * weights, loss.eval(), 3)
79 loss = loss_ops.absolute_difference(self._predictions, self._labels,
82 self.assertAlmostEqual(5.6, loss.eval(), 3
[all...] |
| /external/tensorflow/tensorflow/python/estimator/ |
| model_fn_test.py | 54 loss=constant_op.constant(1.),
60 loss = constant_op.constant(1.)
61 predictions = {'loss': loss}
66 loss=loss,
68 eval_metric_ops={'loss': (control_flow_ops.no_op(), loss)},
79 """Tests that error is raised when loss is a number (not Tensor)."""
81 with self.assertRaisesRegexp(TypeError, 'loss must be Tensor')
[all...] |
| model_fn.py | 53 LOSS_METRIC_KEY = 'loss'
60 'mode', 'predictions', 'loss', 'train_op', 'eval_metric_ops',
72 loss=None,
85 * For `mode == ModeKeys.TRAIN`: required fields are `loss` and `train_op`.
86 * For `mode == ModeKeys.EVAL`: required field is `loss`.
96 loss = ...
101 loss=loss,
112 loss = ...
114 loss = Non
[all...] |
| /external/tensorflow/tensorflow/contrib/gan/python/losses/python/ |
| losses_impl_test.py | 51 loss = self._g_loss_fn(self._discriminator_gen_outputs)
52 self.assertEqual(self._discriminator_gen_outputs.dtype, loss.dtype)
53 self.assertEqual(self._generator_loss_name, loss.op.name)
55 self.assertAlmostEqual(self._expected_g_loss, loss.eval(), 5)
58 loss = self._d_loss_fn(
60 self.assertEqual(self._discriminator_gen_outputs.dtype, loss.dtype)
61 self.assertEqual(self._discriminator_loss_name, loss.op.name)
63 self.assertAlmostEqual(self._expected_d_loss, loss.eval(), 5)
79 loss = self._g_loss_fn(
81 self.assertAllEqual([4], loss.shape
[all...] |
| losses_impl.py | 82 """Wasserstein generator loss for GANs.
93 scope: The scope for the operations performed in computing the loss.
94 loss_collection: collection to which this loss will be added.
95 reduction: A `tf.losses.Reduction` to apply to loss.
96 add_summaries: Whether or not to add detailed summaries for the loss.
99 A loss Tensor. The shape depends on `reduction`.
105 loss = - discriminator_gen_outputs
106 loss = losses.compute_weighted_loss(
107 loss, weights, scope, loss_collection, reduction)
110 summary.scalar('generator_wass_loss', loss)
[all...] |
| /external/tensorflow/tensorflow/contrib/kfac/python/kernel_tests/ |
| loss_functions_test.py | 56 loss = loss_functions.CategoricalLogitsNegativeLogProbLoss(
58 sample = loss.sample(42)
70 loss = loss_functions.CategoricalLogitsNegativeLogProbLoss(
72 neg_log_prob = loss.evaluate()
92 loss = loss_functions.CategoricalLogitsNegativeLogProbLoss(
94 neg_log_prob = loss.evaluate_on_sample(42)
101 """Ensure this loss function supports registering multiple minibatches."""
104 loss = None
109 if loss is None:
110 loss = loss_functions.CategoricalLogitsNegativeLogProbLoss(logits
[all...] |
| /external/tensorflow/tensorflow/python/kernel_tests/ |
| losses_test.py | 54 loss = losses.absolute_difference(self._predictions, self._predictions)
56 self.assertAlmostEqual(0.0, loss.eval(), 3)
59 loss = losses.absolute_difference(self._labels, self._predictions)
61 self.assertAlmostEqual(5.5, loss.eval(), 3)
65 loss = losses.absolute_difference(self._labels, self._predictions, weights)
67 self.assertAlmostEqual(5.5 * weights, loss.eval(), 3)
71 loss = losses.absolute_difference(self._labels, self._predictions,
74 self.assertAlmostEqual(5.5 * weights, loss.eval(), 3)
78 loss = losses.absolute_difference(self._labels, self._predictions, weights)
80 self.assertAlmostEqual(5.6, loss.eval(), 3
[all...] |
| /external/tensorflow/tensorflow/contrib/kernel_methods/python/ |
| losses_test.py | 98 loss = losses.sparse_multiclass_hinge_loss(labels, logits)
100 loss.eval()
103 """Loss is 0 if true class logits sufficiently higher than other classes."""
108 loss = losses.sparse_multiclass_hinge_loss(labels, logits)
109 self.assertAlmostEqual(loss.eval(), 0.0, 3)
112 """Loss is 0 if true class logits sufficiently higher than other classes."""
117 loss = losses.sparse_multiclass_hinge_loss(labels, logits)
118 self.assertAlmostEqual(loss.eval(), 0.0, 3)
137 loss = losses.sparse_multiclass_hinge_loss(labels, logits)
138 result = loss.eval(feed_dict={logits: logits_np, labels: labels_np}
[all...] |
| /external/valgrind/memcheck/tests/ |
| pointer-trace.stderr.exp | 1 1,000 bytes in 1 blocks are definitely lost in loss record ... of ...
|
| static_malloc.stderr.exp | 1 10 bytes in 1 blocks are definitely lost in loss record ... of ...
|
| trivialleak.stderr.exp | 1 1,000 bytes in 1,000 blocks are definitely lost in loss record ... of ...
|
| leak-delta.stderr.exp | 2 10 bytes in 1 blocks are still reachable in loss record ... of ...
9 10 (+10) bytes in 1 (+1) blocks are definitely lost in loss record ... of ...
14 21 (+21) bytes in 1 (+1) blocks are still reachable in loss record ... of ...
20 65 (+65) bytes in 2 (+2) blocks are still reachable in loss record ... of ...
27 10 (+10) bytes in 1 (+1) blocks are still reachable in loss record ... of ...
33 0 (-10) bytes in 0 (-1) blocks are still reachable in loss record ... of ...
38 10 (+10) bytes in 1 (+1) blocks are definitely lost in loss record ... of ...
44 0 (-10) bytes in 0 (-1) blocks are definitely lost in loss record ... of ...
49 10 (+10) bytes in 1 (+1) blocks are still reachable in loss record ... of ...
55 32 (+32) bytes in 1 (+1) blocks are definitely lost in loss record ... of ..
[all...] |
| /external/valgrind/gdbserver_tests/ |
| mcleak.stderrB.exp | 3 10 bytes in 1 blocks are still reachable in loss record ... of ...
7 10 (+10) bytes in 1 (+1) blocks are definitely lost in loss record ... of ...
11 21 (+21) bytes in 1 (+1) blocks are still reachable in loss record ... of ...
15 65 (+65) bytes in 2 (+2) blocks are still reachable in loss record ... of ...
19 10 (+10) bytes in 1 (+1) blocks are still reachable in loss record ... of ...
23 0 (-10) bytes in 0 (-1) blocks are still reachable in loss record ... of ...
27 10 (+10) bytes in 1 (+1) blocks are definitely lost in loss record ... of ...
31 0 (-10) bytes in 0 (-1) blocks are definitely lost in loss record ... of ...
35 10 (+10) bytes in 1 (+1) blocks are still reachable in loss record ... of ...
39 32 (+32) bytes in 1 (+1) blocks are definitely lost in loss record ... of ..
[all...] |
| /external/tensorflow/tensorflow/contrib/kfac/examples/ |
| mlp.py | 79 loss: 0-D Tensor representing loss to be minimized.
88 loss = tf.reduce_mean(
105 return loss, accuracy
108 def minimize(loss, accuracy, layer_collection, session_config=None):
109 """Minimize 'loss' with KfacOptimizer.
112 loss: 0-D Tensor. Loss to be minimized.
131 train_op = optimizer.minimize(loss, global_step=global_step)
145 [global_step, loss, accuracy, train_op, optimizer.cov_update_op]
[all...] |
| convnet.py | 143 loss: 0-D Tensor representing loss to be minimized.
158 loss = tf.reduce_mean(
164 tf.summary.scalar("loss", loss)
178 return loss, accuracy
181 def minimize_loss_single_machine(loss,
185 """Minimize loss with K-FAC on a single machine.
190 loss: 0-D Tensor. Loss to be minimized
[all...] |
| /external/tensorflow/tensorflow/contrib/layers/python/layers/ |
| regularizers_test.py | 78 loss = regularizers.l1_l2_regularizer(1.0, 1.0)(tensor)
79 self.assertEquals(loss.op.name, 'l1_l2_regularizer')
80 self.assertAlmostEqual(loss.eval(), num_elem + num_elem / 2, 5)
86 loss = regularizers.l1_l2_regularizer(0.0, 1.0)(tensor)
88 self.assertEquals(loss.op.name, 'l1_l2_regularizer')
89 self.assertAlmostEqual(loss.eval(), num_elem / 2, 5)
95 loss = regularizers.l1_l2_regularizer(1.0, 0.0)(tensor)
97 self.assertEquals(loss.op.name, 'l1_l2_regularizer')
98 self.assertAlmostEqual(loss.eval(), num_elem, 5)
103 loss = regularizers.l1_l2_regularizer(0.0, 0.0)(tensor
[all...] |
| target_column_test.py | 38 5. / 3, sess.run(target_column.loss(prediction, labels, {})))
49 sess.run(target_column.loss(prediction, labels, features)),
68 sess.run(target_column.loss(logits, labels, {})),
82 sess.run(target_column.loss(logits, labels, features)),
106 sess.run(target_column.loss(logits, labels, {})))
118 1.5514446, sess.run(target_column.loss(logits, labels, features)))
134 loss_op, update_op = eval_dict["loss"]
146 loss = target_column.loss(predictions, labels, {})
148 # < 0) but it is within the [-1,1] margin. There is a 0.5 loss incurred b
[all...] |
| /external/tensorflow/tensorflow/python/ops/ |
| nn_xent_test.py | 59 loss = nn_impl.sigmoid_cross_entropy_with_logits(
61 self.assertEqual("mylogistic", loss.op.name)
68 loss = nn_impl.sigmoid_cross_entropy_with_logits(
71 tf_loss = loss.eval()
79 loss = nn_impl.sigmoid_cross_entropy_with_logits(
82 tf_loss = loss.eval()
89 loss = nn_impl.sigmoid_cross_entropy_with_logits(
91 err = gradient_checker.compute_gradient_error(logits, sizes, loss, sizes)
92 print("logistic loss gradient err = ", err)
99 loss = nn_impl.sigmoid_cross_entropy_with_logits
[all...] |
| /external/tensorflow/tensorflow/contrib/learn/python/learn/estimators/ |
| model_fn.py | 64 'predictions', 'loss', 'train_op', 'eval_metric_ops',
73 loss=None,
102 loss: Training loss `Tensor`.
130 get_graph_from_inputs((predictions, loss, train_op))
140 # Validate loss.
141 if loss is None:
143 raise ValueError('Missing loss.')
145 loss = ops.convert_to_tensor(loss)
[all...] |
| /external/tensorflow/tensorflow/core/kernels/ |
| sparse_xent_op_gpu.cu.cc | 36 typename TTypes<T>::Vec scratch, typename TTypes<T>::Vec loss,
39 scratch, loss, backprop);
|
| xent_op_gpu.cu.cc | 37 typename TTypes<T>::Vec loss,
39 XentEigenImpl<GPUDevice, T>::Compute(d, logits, labels, scratch, loss,
|
| /prebuilts/go/darwin-x86/test/fixedbugs/ |
| issue12677.go | 7 // Issue 12677: Type loss during export/import of inlined function body.
|
| /prebuilts/go/linux-x86/test/fixedbugs/ |
| issue12677.go | 7 // Issue 12677: Type loss during export/import of inlined function body.
|
| /external/tensorflow/tensorflow/contrib/eager/python/examples/mnist/ |
| mnist_graph_test.py | 49 # Define the loss tensor (as opposed to a loss function when
53 loss = mnist.loss(predictions, labels)
55 train_op = optimizer.minimize(loss)
|
