/external/tensorflow/tensorflow/python/keras/ |
metrics_functional_test.py | 43 y_pred = K.variable(np.random.random((6, 7))) 44 self.assertEqual(K.eval(metric(y_true, y_pred)).shape, (6,)) 48 y_pred = K.variable([[0.8, 0.2], [0.6, 0.4], [0.7, 0.3], [0.9, 0.1]]) 49 print(K.eval(metric(y_true, y_pred))) 50 self.assertAllEqual(K.eval(metric(y_true, y_pred)), [0., 1., 1., 1.]) 54 y_pred = K.variable([[0.8, 0.2], [0.6, 0.4], [0.7, 0.3], [0.9, 0.1]]) 55 print(K.eval(metric(y_true, y_pred))) 56 self.assertAllEqual(K.eval(metric(y_true, y_pred)), [0., 1., 1., 1.]) 62 y_pred = K.variable(np.random.random((6, 7))) 63 self.assertEqual(K.eval(metric(y_true, y_pred)).shape, (6,) [all...] |
losses_test.py | 59 def __call__(self, y_true, y_pred, sample_weight=None): 60 return (self.mse_fraction * keras.losses.mse(y_true, y_pred) + 61 (1 - self.mse_fraction) * keras.losses.mae(y_true, y_pred)) 143 y_pred = keras.backend.variable(np.array([[0.3, 0.2, 0.1], 147 loss = keras.backend.eval(keras.losses.categorical_hinge(y_true, y_pred)) 192 y_pred = constant_op.constant([[4., 8.], [12., 3.]]) 194 loss = mse_obj(y_true, y_pred, sample_weight=sample_weight) 221 y_pred = constant_op.constant([4, 8, 12, 8, 1, 3], 224 loss = mse_obj(y_true, y_pred) 230 y_pred = constant_op.constant([4, 8, 12, 8, 1, 3] [all...] |
metrics_confusion_matrix_test.py | 58 y_pred = constant_op.constant(((0, 0, 1, 1, 0), (1, 1, 1, 1, 1), 61 update_op = fp_obj.update_state(y_true, y_pred) 71 y_pred = constant_op.constant(((0, 0, 1, 1, 0), (1, 1, 1, 1, 1), 74 result = fp_obj(y_true, y_pred, sample_weight=sample_weight) 81 y_pred = constant_op.constant(((0.9, 0.2, 0.8, 0.1), (0.2, 0.9, 0.7, 0.6), 86 update_op = fp_obj.update_state(y_true, y_pred) 95 y_pred = constant_op.constant(((0.9, 0.2, 0.8, 0.1), (0.2, 0.9, 0.7, 0.6), 102 result = fp_obj(y_true, y_pred, sample_weight=sample_weight) 138 y_pred = constant_op.constant(((0, 0, 1, 1, 0), (1, 1, 1, 1, 1), 141 update_op = fn_obj.update_state(y_true, y_pred) [all...] |
metrics.py | 118 def update_state(self, y_true, y_pred, sample_weight=None): 120 y_pred = tf.cast(y_pred, tf.bool) 122 values = tf.logical_and(tf.equal(y_true, True), tf.equal(y_pred, True)) 450 # metric = mean(|y_pred - y_true| / normalizer) 479 def update_state(self, y_true, y_pred, sample_weight=None): 484 y_pred: The predicted values. 493 y_pred = math_ops.cast(y_pred, self._dtype) 494 y_pred, y_true, sample_weight = squeeze_or_expand_dimensions [all...] |
losses.py | 45 * `call()`: Contains the logic for loss calculation using `y_true`, `y_pred`. 50 def call(self, y_true, y_pred): 51 y_pred = ops.convert_to_tensor(y_pred) 52 y_true = math_ops.cast(y_true, y_pred.dtype) 53 return K.mean(math_ops.square(y_pred - y_true), axis=-1) 68 def __call__(self, y_true, y_pred, sample_weight=None): 73 y_pred: The predicted values. 80 the shape of `sample_weight` matches the shape of `y_pred`, then the 81 loss of each measurable element of `y_pred` is scaled by th [all...] |
metrics_test.py | 345 # check y_pred squeeze 470 y_pred = self.l2_norm(self.np_y_pred, axis) 471 self.expected_loss = np.sum(np.multiply(y_true, y_pred), axis=(axis,)) 474 self.y_pred = constant_op.constant(self.np_y_pred) 491 loss = cosine_obj(self.y_true, self.y_pred) 502 self.y_pred, 512 loss = cosine_obj(self.y_true, self.y_pred) 535 y_pred = constant_op.constant(((0, 0, 1, 1, 0), (1, 1, 1, 1, 1), 538 update_op = mae_obj.update_state(y_true, y_pred) 548 y_pred = constant_op.constant(((0, 0, 1, 1, 0), (1, 1, 1, 1, 1) [all...] |
backend.py | [all...] |
/external/tensorflow/tensorflow/python/keras/utils/ |
losses_utils.py | 60 def squeeze_or_expand_dimensions(y_pred, y_true, sample_weight): 63 1. Squeezes last dim of `y_pred` or `y_true` if their rank differs by 1 66 from the new rank of `y_pred`. 73 y_pred: Predicted values, a `Tensor` of arbitrary dimensions. 74 y_true: Optional label `Tensor` whose dimensions match `y_pred`. 76 `y_pred`. 79 Tuple of `y_pred`, `y_true` and `sample_weight`. Each of them possibly has 83 y_pred_shape = y_pred.get_shape() 87 # If sparse matrix is provided as `y_true`, the last dimension in `y_pred` 89 # y_pred = [[.9, .05, .05], [.5, .89, .6], [.05, .01, .94]] (shape=(3, 3) [all...] |
metrics_utils.py | 214 y_pred, 221 For every pair of values in y_true and y_pred: 223 true_positive: y_true == True and y_pred > thresholds 224 false_negatives: y_true == True and y_pred <= thresholds 225 true_negatives: y_true == False and y_pred <= thresholds 226 false_positive: y_true == False and y_pred > thresholds 240 y_true: A `Tensor` whose shape matches `y_pred`. Will be cast to `bool`. 241 y_pred: A floating point `Tensor` of arbitrary shape and whose values are in 257 ValueError: If `y_pred` and `y_true` have mismatched shapes, or if 258 `sample_weight` is not `None` and its shape doesn't match `y_pred`, or i [all...] |
tf_utils_test.py | 137 def custom_loss(y_obs, y_pred): 139 obtained_prediction_box[0] = y_pred 140 return y_pred
|
/external/tensorflow/tensorflow/contrib/learn/python/learn/estimators/ |
_sklearn.py | 157 def _accuracy_score(y_true, y_pred): 158 score = y_true == y_pred 162 def _mean_squared_error(y_true, y_pred): 165 if len(y_pred.shape) > 1: 166 y_pred = np.squeeze(y_pred) 167 return np.average((y_true - y_pred)**2)
|
/external/tensorflow/tensorflow/python/keras/engine/ |
training_gpu_test.py | 47 loss = lambda y_true, y_pred: K.sparse_categorical_crossentropy( # pylint: disable=g-long-lambda 48 y_true, y_pred, axis=axis) 52 loss = lambda y_true, y_pred: K.categorical_crossentropy( # pylint: disable=g-long-lambda 53 y_true, y_pred, axis=axis) 57 loss = lambda y_true, y_pred: K.binary_crossentropy(y_true, y_pred) # pylint: disable=unnecessary-lambda
|
training_utils.py | 815 def call_metric_function(metric_fn, y_true, y_pred, weights=None, mask=None): 818 return metric_fn(y_true, y_pred, sample_weight=weights) 820 mask = math_ops.cast(mask, y_pred.dtype) 823 return metric_fn(y_true, y_pred, sample_weight=mask) 828 return metric_fn(y_true, y_pred, sample_weight=weights) [all...] |
training.py | [all...] |
base_layer.py | [all...] |
/external/libopus/scripts/ |
dump_rnn.py | 32 def binary_crossentrop2(y_true, y_pred): 33 return K.mean(2*K.abs(y_true-0.5) * K.binary_crossentropy(y_pred, y_true), axis=-1)
|
rnn_train.py | 19 def binary_crossentrop2(y_true, y_pred): 20 return K.mean(2*K.abs(y_true-0.5) * K.binary_crossentropy(y_pred, y_true), axis=-1)
|
/external/tensorflow/tensorflow/contrib/seq2seq/python/ops/ |
loss.py | 160 def __call__(self, y_true, y_pred, sample_weight=None): 162 return sequence_loss(y_pred, y_true, sample_weight, 170 def call(self, y_true, y_pred):
|
/external/tensorflow/tensorflow/contrib/losses/python/metric_learning/ |
metric_loss_ops_test.py | 428 # y_pred = self._get_cluster_ics(D, medoid_ics) 431 y_pred = cluster_ics 433 if sum(y_pred == cluster_idx) == 0: 439 pdists[medoid_ics[cluster_idx], y_pred == cluster_idx]) + 441 y_gt, y_pred))) 443 pdist_in = pdists[y_pred == cluster_idx, :] 444 pdist_in = pdist_in[:, y_pred == cluster_idx] 448 for i in range(y_pred.size): 449 if y_pred[i] != cluster_idx: 465 y_pred == cluster_idx)[0][max_score_idx [all...] |
/external/tensorflow/tensorflow/python/keras/mixed_precision/experimental/ |
keras_test.py | 280 def loss_fn(y_true, y_pred): 282 return math_ops.reduce_mean(y_pred) 358 def loss_fn(y_true, y_pred): 360 self.assertEqual(y_pred.dtype, dtypes.float32) 361 return math_ops.reduce_mean(y_pred)
|
/external/tensorflow/tensorflow/python/keras/layers/ |
local_test.py | 359 def xent(y_true, y_pred): 366 logits=y_pred)
|
/external/libaom/libaom/av1/encoder/ |
temporal_filter.c | 255 const uint8_t *y_frame1, int y_stride, const uint8_t *y_pred, 279 calculate_squared_errors(y_frame1, y_stride, y_pred, y_buf_stride, y_diff_sse, 288 const int pixel_value = y_pred[i * y_buf_stride + j]; 416 const uint16_t *y_pred = CONVERT_TO_SHORTPTR(yp); local 427 highbd_calculate_squared_errors(y_frame1, y_stride, y_pred, y_buf_stride, 436 const int pixel_value = y_pred[i * y_buf_stride + j]; [all...] |
/external/libvpx/libvpx/vp9/encoder/ |
vp9_temporal_filter.c | 224 const uint8_t *y_frame1, int y_stride, const uint8_t *y_pred, 254 y_frame1[i * (int)y_stride + j] - y_pred[i * (int)block_width + j]; 273 const int pixel_value = y_pred[i * y_buf_stride + j]; [all...] |
/external/tensorflow/tensorflow/contrib/distribute/python/ |
keras_test.py | [all...] |
/external/tensorflow/tensorflow/python/keras/saving/ |
hdf5_format_test.py | 471 def custom_loss(y_true, y_pred): 472 return keras.losses.mse(y_true, y_pred) [all...] |