/external/tensorflow/tensorflow/lite/models/smartreply/ |
predictor_test.cc | 71 std::vector<PredictorResponse> predictions; local 73 GetSegmentPredictions({"Welcome"}, *model_, /*config=*/{{}}, &predictions); 74 EXPECT_GT(predictions.size(), 0); 77 for (const auto &item : predictions) { 85 &predictions, 90 std::vector<PredictorResponse> predictions; local 93 &predictions); 94 EXPECT_GT(predictions.size(), 0); 97 for (const auto &item : predictions) { 104 EXPECT_THAT(&predictions, IncludeAnyResponesIn(std::unordered_set<string> 109 std::vector<PredictorResponse> predictions; local 137 std::vector<PredictorResponse> predictions; local [all...] |
/external/tensorflow/tensorflow/python/kernel_tests/ |
in_topk_op_test.py | 31 def _validateInTopK(self, predictions, target, k, expected): 34 precision = nn_ops.in_top_k(predictions, target, k) 40 predictions = [[0.1, 0.3, 0.2, 0.4], [0.1, 0.2, 0.3, 0.4]] 42 self._validateInTopK(predictions, target, 1, [True, False]) 45 predictions = [[0.1, 0.3, 0.2, 0.4], [0.1, 0.2, 0.3, 0.4]] 47 self._validateInTopK(predictions, target, 2, [False, True]) 51 predictions = [[0.1, 0.3, 0.2, 0.2], [0.1, 0.3, 0.2, 0.2]] 53 self._validateInTopK(predictions, target, 2, [True, True]) 56 predictions = [[0.1, 0.3, 0.2, 0.4], [0.1, 0.2, 0.3, 0.4]] 58 self._validateInTopK(predictions, target, 2, [False, True] [all...] |
metrics_test.py | 563 predictions=array_ops.ones((10, 1)), 573 predictions=array_ops.ones((10, 1)), 582 predictions=array_ops.ones((10, 1)), 589 predictions = array_ops.ones((10, 3)) 592 metrics.accuracy(labels, predictions) 596 predictions = array_ops.ones((10, 3)) 600 metrics.accuracy(labels, predictions, weights) 604 predictions = random_ops.random_uniform( 608 accuracy, update_op = metrics.accuracy(labels, predictions) 625 # Create the queue that populates the predictions [all...] |
/external/tensorflow/tensorflow/python/ops/ |
confusion_matrix.py | 34 labels, predictions, expected_rank_diff=0, name=None): 41 But, for example, if `labels` contains class IDs and `predictions` contains 1 42 probability per class, we expect `predictions` to have 1 more dimension than 44 `labels` if `rank(predictions) - rank(labels) == 0`, and 45 `predictions` if `rank(predictions) - rank(labels) == 2`. 51 labels: Label values, a `Tensor` whose dimensions match `predictions`. 52 predictions: Predicted values, a `Tensor` of arbitrary dimensions. 53 expected_rank_diff: Expected result of `rank(predictions) - rank(labels)`. 57 Tuple of `labels` and `predictions`, possibly with last dim squeezed [all...] |
metrics_impl.py | 88 def _remove_squeezable_dimensions(predictions, labels, weights): 91 Squeezes last dim of `predictions` or `labels` if their rank differs by 1 94 new rank of `predictions`. 102 predictions: Predicted values, a `Tensor` of arbitrary dimensions. 103 labels: Optional label `Tensor` whose dimensions match `predictions`. 105 `predictions`. 108 Tuple of `predictions`, `labels` and `weights`. Each of them possibly has 111 predictions = ops.convert_to_tensor(predictions) 113 labels, predictions = confusion_matrix.remove_squeezable_dimensions [all...] |
/external/tensorflow/tensorflow/contrib/learn/python/learn/estimators/ |
model_fn_test.py | 47 def create_model_fn_ops(self, predictions, output_alternatives, 52 predictions=predictions, 70 self.assertEqual(model_fn_ops.predictions, estimator_spec.predictions) 82 predictions = self.create_predictions() 84 predictions, None, mode=model_fn.ModeKeys.INFER) 90 predictions = self.create_predictions() 92 constants.ProblemType.LINEAR_REGRESSION, predictions)} 94 predictions, output_alternatives, mode=model_fn.ModeKeys.INFER [all...] |
logistic_regressor.py | 43 `(features, labels, mode) -> (predictions, loss, train_op)`. 44 Expects the returned predictions to be probabilities in [0.0, 1.0]. 54 predictions, loss, train_op = model_fn(features, labels, mode) 58 predictions=predictions, 64 predictions=predictions, 70 'predictions': predictions 109 `(features, labels, mode) -> (predictions, loss, train_op)` [all...] |
model_fn.py | 72 'predictions', 'loss', 'train_op', 'eval_metric_ops', 88 predictions=None, 98 For a multi-headed model, the predictions dict here will contain the outputs 117 predictions: Predictions `Tensor` or dict of `Tensor`. 146 get_graph_from_inputs((predictions, loss, train_op)) 168 # Validate predictions. 169 if predictions is None: 171 raise ValueError('Missing predictions.') 173 if isinstance(predictions, dict) [all...] |
/external/tensorflow/tensorflow/contrib/metrics/python/ops/ |
confusion_matrix_ops.py | 25 def confusion_matrix(labels, predictions, num_classes=None, dtype=dtypes.int32, 28 return cm.confusion_matrix(labels=labels, predictions=predictions,
|
metric_ops_test.py | 487 predictions=array_ops.ones((10, 1)), 496 predictions=array_ops.ones((10, 1)), 504 predictions=array_ops.ones((10, 1)), 510 predictions = array_ops.ones((10, 3)) 513 metrics.streaming_accuracy(predictions, labels) 516 predictions = array_ops.ones((10, 3)) 520 metrics.streaming_accuracy(predictions, labels, weights) 523 predictions = random_ops.random_uniform( 527 accuracy, update_op = metrics.streaming_accuracy(predictions, labels) 543 # Create the queue that populates the predictions [all...] |
metric_ops.py | 49 'order of the labels and predictions arguments has been switched.') 50 def streaming_true_positives(predictions, 61 predictions: The predicted values, a `Tensor` of arbitrary dimensions. Will 64 `predictions`. Will be cast to `bool`. 80 ValueError: If `predictions` and `labels` have mismatched shapes, or if 81 `weights` is not `None` and its shape doesn't match `predictions`, or if 86 predictions=predictions, 95 'order of the labels and predictions arguments has been switched.') 96 def streaming_true_negatives(predictions, [all...] |
/external/tensorflow/tensorflow/contrib/learn/python/learn/ |
metric_spec_test.py | 35 def _fn0(predictions, labels, weights=None): 36 self.assertEqual("p1_value", predictions) 41 def _fn1(predictions, targets, weights=None): 42 self.assertEqual("p1_value", predictions) 150 def _fn(predictions): 151 self.assertEqual(predictions_, predictions) 191 def _fn0(predictions, labels): 192 self.assertEqual("p1_value", predictions) 196 def _fn1(predictions, targets): 197 self.assertEqual("p1_value", predictions) [all...] |
metric_spec.py | 37 '`labels`, `predictions`, and optionally `weights`.') 59 _CANONICAL_PREDICTIONS_ARG = 'predictions' 116 This returns a function that takes only named args `labels`, `predictions`, 119 passed (usually by name, but positionally if both it and `predictions` need 122 passed by name. Otherwise, `predictions` are passed positionally as the 134 Function accepting only named args `labels, `predictions`, and `weights`, 160 # Both labels and predictions are named args. 162 _sentinel=None, labels=None, predictions=None, weights=None): 166 predictions_arg: predictions, 174 # labels is a named arg, and first. predictions is not a named arg, so w [all...] |
/external/tensorflow/tensorflow/contrib/learn/python/learn/utils/ |
export.py | 95 def generic_signature_fn(examples, unused_features, predictions): 96 """Creates generic signature from given examples and predictions. 104 predictions: `Tensor` or `dict` of `Tensor`s. 116 if not isinstance(predictions, dict): 117 predictions = {'outputs': predictions} 118 tensors.update(predictions) 127 def classification_signature_fn(examples, unused_features, predictions): 128 """Creates classification signature from given examples and predictions. 133 predictions: `Tensor` or dict of tensors that contains the classes tenso [all...] |
/external/tensorflow/tensorflow/python/ops/losses/ |
losses_impl.py | 213 labels, predictions, weights=1.0, scope=None, 222 `weights` matches the shape of `predictions`, then the loss of each 223 measurable element of `predictions` is scaled by the corresponding value of 227 labels: The ground truth output tensor, same dimensions as 'predictions'. 228 predictions: The predicted outputs. 241 ValueError: If the shape of `predictions` doesn't match that of 243 or `predictions` is None. 252 if predictions is None: 253 raise ValueError("predictions must not be None.") 255 (predictions, labels, weights)) as scope [all...] |
/external/tensorflow/tensorflow/contrib/learn/python/learn/ops/ |
losses_ops.py | 41 predictions = nn.xw_plus_b(tensor_in, weights, biases) 42 if len(labels.get_shape()) == 1 and len(predictions.get_shape()) == 2: 43 predictions = array_ops_.squeeze(predictions, axis=[1]) 44 return predictions, losses.mean_squared_error(labels, predictions) 58 predictions, use `tf.argmax` on the returned probabilities. 74 `tuple` of softmax predictions and loss `Tensor`s.
|
/external/tensorflow/tensorflow/contrib/tensor_forest/client/ |
eval_metrics.py | 46 def _accuracy(predictions, targets, weights=None): 48 labels=targets, predictions=predictions, weights=weights) 83 def _predictions(predictions, unused_targets, **unused_kwargs): 84 return predictions 95 def _precision(predictions, targets, weights=None): 97 labels=targets, predictions=predictions, weights=weights) 100 def _precision_at_thresholds(predictions, targets, weights=None): 103 predictions=array_ops.slice(predictions, [0, 1], [-1, 1]) [all...] |
/external/tensorflow/tensorflow/contrib/losses/python/losses/ |
loss_ops.py | 240 def absolute_difference(predictions, labels=None, weights=1.0, scope=None): 247 `weights` matches the shape of `predictions`, then the loss of each 248 measurable element of `predictions` is scaled by the corresponding value of 252 predictions: The predicted outputs. 253 labels: The ground truth output tensor, same dimensions as 'predictions'. 255 [batch_size] or a tensor whose shape matches `predictions`. 262 ValueError: If the shape of `predictions` doesn't match that of `labels` or 266 [predictions, labels, weights]) as scope: 267 predictions.get_shape().assert_is_compatible_with(labels.get_shape()) 268 predictions = math_ops.cast(predictions, dtypes.float32 [all...] |
/external/tensorflow/tensorflow/contrib/metrics/python/metrics/ |
classification.py | 32 def accuracy(predictions, labels, weights=None, name=None): 33 """Computes the percentage of times that predictions matches labels. 36 predictions: the predicted values, a `Tensor` whose dtype and shape 55 if not labels.dtype.is_compatible_with(predictions.dtype): 56 raise ValueError('Dtypes of predictions and labels should match. ' 57 'Given: predictions (%r) and labels (%r)' % 58 (predictions.dtype, labels.dtype)) 59 with ops.name_scope(name, 'accuracy', values=[predictions, labels]): 61 math_ops.equal(predictions, labels), dtypes.float32) 70 def f1_score(labels, predictions, weights=None, num_thresholds=200 [all...] |
classification_test.py | 126 predictions=array_ops.ones((10, 1)), 142 predictions=array_ops.ones((10, 1)), 151 predictions=array_ops.ones((10, 1)), 158 predictions = random_ops.random_uniform( 162 f1, f1_op = classification.f1_score(predictions, labels, num_thresholds=3) 180 predictions = constant_op.constant(inputs, dtype=dtypes.float32) 182 f1, f1_op = classification.f1_score(predictions, labels, num_thresholds=3) 190 predictions = constant_op.constant( 193 f1, f1_op = classification.f1_score(predictions, labels, num_thresholds=1) 205 predictions = constant_op.constant(inputs, dtype=dtypes.float32 [all...] |
/external/tensorflow/tensorflow/contrib/distribute/python/ |
metrics_v1_test.py | 33 # First four batches of x: labels, predictions -> (labels == predictions) 39 lambda x: {"labels": x % 5, "predictions": x % 3}).batch( 44 # First four batches of labels, predictions: {TP, FP, TN, FN} 52 "predictions": [True, True, False, False]}).repeat().batch( 57 # First four batches of labels, predictions: {TP, FP, TN, FN} 65 "predictions": [1.0, 0.75, 0.25, 0.]}).repeat().batch( 72 "predictions": [1., .75, .25, 0.]}).repeat() 153 predictions = x["predictions"] [all...] |
/external/tensorflow/tensorflow/contrib/eager/python/ |
metrics_impl.py | 366 """Calculates how often `predictions` matches `labels`. 376 def call(self, labels, predictions, weights=None): 379 For example, if labels is [1, 2, 3, 4] and predictions is [0, 2, 3, 4] 383 `labels` and `predictions` should have the same shape and type. 388 predictions: Tensor with the predicted label for each example. 395 array_ops.shape(labels), array_ops.shape(predictions), 396 message="Shapes of labels and predictions are unequal") 397 matches = math_ops.equal(labels, predictions) 401 return labels, predictions 402 return labels, predictions, weight [all...] |
/external/tensorflow/tensorflow/contrib/linear_optimizer/python/kernel_tests/ |
sdca_ops_test.py | 208 def get_binary_predictions_for_logistic(predictions, cutoff=0.5): 210 math_ops.greater_equal(predictions, 211 array_ops.ones_like(predictions) * cutoff), 215 def get_binary_predictions_for_hinge(predictions): 217 math_ops.greater_equal(predictions, array_ops.zeros_like(predictions)), 262 predictions = lr.predictions(examples) 278 predicted_labels = get_binary_predictions_for_logistic(predictions) 310 predictions = lr.predictions(examples [all...] |
/external/tensorflow/tensorflow/contrib/boosted_trees/python/utils/ |
losses.py | 30 def per_example_squared_hinge_loss(labels, weights, predictions): 31 loss = losses.hinge_loss(labels=labels, logits=predictions, weights=weights) 35 def per_example_logistic_loss(labels, weights, predictions): 36 """Logistic loss given labels, example weights and predictions. 41 predictions: Rank 2 (N, 1) tensor of per-example predictions. 49 labels=labels, logits=predictions) 60 def per_example_quantile_regression_loss(labels, weights, predictions, 71 predictions: Rank 2 (N, D) tensor of per-example predictions [all...] |
/external/tensorflow/tensorflow/contrib/distribute/python/examples/ |
simple_estimator_example.py | 42 predictions = {"logits": logits} 43 return tf.estimator.EstimatorSpec(mode, predictions=predictions) 97 predictions = [prediction_iterable.next() for _ in range(10)] 98 print("Prediction results: {}".format(predictions))
|