/external/tensorflow/tensorflow/contrib/estimator/python/estimator/ |
H A D | dnn.py | 93 dropout=None, 114 dropout: When not `None`, the probability we will drop out a given 130 dropout=dropout,
|
H A D | dnn_linear_combined_test.py | 50 dropout=None, 61 dnn_dropout=dropout,
|
/external/tensorflow/tensorflow/python/estimator/canned/ |
H A D | dnn.py | 49 dropout, input_layer_partitioner): 59 dropout: When not `None`, the probability we will drop out a given 102 if dropout is not None and mode == model_fn.ModeKeys.TRAIN: 103 net = core_layers.dropout(net, rate=dropout, training=True) 128 dropout=None, 146 dropout: When not `None`, the probability we will drop out a given 182 dropout=dropout, 279 dropout [all...] |
/external/tensorflow/tensorflow/contrib/cudnn_rnn/python/ops/ |
H A D | cudnn_rnn_ops.py | 812 dropout=0., 837 dropout: whether to enable dropout. With it is 0, dropout is disabled. 838 seed: the op seed used for initializing dropout. See @{tf.set_random_seed} 857 dropout=dropout, 871 dropout=0., 895 dropout: whether to enable dropout [all...] |
/external/tensorflow/tensorflow/contrib/slim/python/slim/nets/ |
H A D | vgg.py | 89 dropout_keep_prob: the probability that activations are kept in the dropout 117 net = layers_lib.dropout( 120 net = layers_lib.dropout( 154 dropout_keep_prob: the probability that activations are kept in the dropout 182 net = layers_lib.dropout( 185 net = layers_lib.dropout( 219 dropout_keep_prob: the probability that activations are kept in the dropout 247 net = layers_lib.dropout( 250 net = layers_lib.dropout(
|
H A D | alexnet.py | 86 dropout_keep_prob: the probability that activations are kept in the dropout 117 net = layers_lib.dropout( 120 net = layers_lib.dropout(
|
H A D | overfeat.py | 82 dropout_keep_prob: the probability that activations are kept in the dropout 113 net = layers_lib.dropout( 116 net = layers_lib.dropout(
|
/external/tensorflow/tensorflow/contrib/learn/python/learn/estimators/ |
H A D | dnn.py | 103 * dropout: When not `None`, the probability we will drop out a given 124 dropout = params.get("dropout") 173 if dropout is not None and mode == model_fn.ModeKeys.TRAIN: 174 net = layers.dropout(net, keep_prob=(1.0 - dropout)) 300 dropout=None, 332 dropout: When not `None`, the probability we will drop out a given 374 "dropout": dropout, [all...] |
H A D | nonlinear_test.py | 101 dropout=0.0, 113 dropout=0.1, 125 dropout=0.9,
|
H A D | composable_model.py | 264 dropout=None, 279 dropout: When not None, the probability we will drop out 301 self._dropout = dropout 374 net = layers.dropout(net, keep_prob=(1.0 - self._dropout))
|
/external/tensorflow/tensorflow/python/ops/ |
H A D | nn_test.py | 277 # Runs dropout with 0-1 tensor 10 times, sum the number of ones and validate 287 dropout = nn_ops.dropout(t, keep_prob) 289 self.assertEqual([x_dim, y_dim], dropout.get_shape()) 291 value = dropout.eval() 304 # Runs dropout with 0-1 tensor 10 times, sum the number of ones and validate 315 dropout = nn_ops.dropout(t, keep_prob, noise_shape=[x_dim, 1]) 316 self.assertEqual([x_dim, y_dim], dropout.get_shape()) 319 value = dropout [all...] |
/external/tensorflow/tensorflow/examples/tutorials/layers/ |
H A D | cnn_mnist.py | 80 # Add dropout operation; 0.6 probability that element will be kept 81 dropout = tf.layers.dropout( 87 logits = tf.layers.dense(inputs=dropout, units=10)
|
/external/tensorflow/tensorflow/examples/speech_commands/ |
H A D | models.py | 79 placeholder node that can be used to control the dropout amount. 92 TensorFlow node outputting logits results, and optionally a dropout 148 TensorFlow node outputting logits results, and optionally a dropout 201 During training, dropout nodes are introduced after each relu, controlled by a 210 TensorFlow node outputting logits results, and optionally a dropout 231 first_dropout = tf.nn.dropout(first_relu, dropout_prob) 250 second_dropout = tf.nn.dropout(second_relu, dropout_prob) 307 During training, dropout nodes are introduced after the relu, controlled by a 316 TensorFlow node outputting logits results, and optionally a dropout 340 first_dropout = tf.nn.dropout(first_rel [all...] |
/external/tensorflow/tensorflow/python/keras/_impl/keras/layers/ |
H A D | recurrent.py | 833 dropout: Float between 0 and 1. 854 dropout=0., 874 self.dropout = min(1., max(0., dropout)) 907 if 0 < self.dropout < 1 and self._dropout_mask is None: 911 self.dropout, 937 if 0 < self.dropout + self.recurrent_dropout: 968 'dropout': 969 self.dropout, 1006 dropout 1137 def dropout(self): member in class:SimpleRNN 1622 def dropout(self): member in class:GRU 2157 def dropout(self): member in class:LSTM [all...] |
H A D | convolutional_recurrent.py | 261 dropout: Float between 0 and 1. 323 dropout=0., 355 self.dropout = min(1., max(0., dropout)) 485 if self.implementation == 0 and 0 < self.dropout < 1: 491 return K.dropout(ones, self.dropout) 510 return K.dropout(ones, self.recurrent_dropout) 595 'dropout': 596 self.dropout, [all...] |
H A D | core_test.py | 52 dropout = keras.layers.Dropout(0.5) 53 self.assertEqual(True, dropout.supports_masking)
|
/external/tensorflow/tensorflow/contrib/cudnn_rnn/python/kernel_tests/ |
H A D | cudnn_rnn_ops_test.py | 66 dropout=0.): 84 dropout=dropout) 420 batch_size, seq_length, dir_count, dropout, 431 dropout=dropout) 522 # Cudnn scales result for dropout during training, therefore dropout has no 525 # demonstrative of the dropout-invariant nature of CudnnRnn.) 527 for (config, dropout) i [all...] |
/external/tensorflow/tensorflow/contrib/cudnn_rnn/python/layers/ |
H A D | cudnn_rnn.py | 158 dropout=0., 179 dropout: dropout rate, a number between [0, 1]. Dropout is applied between 180 each layer (no dropout is applied for a model with a single layer). 181 When set to 0, dropout is disabled. 182 seed: the op seed used for initializing dropout. See @{tf.set_random_seed} 208 self._dropout = dropout 464 dropout=self._dropout, 477 dropout=self._dropout,
|
/external/tensorflow/tensorflow/contrib/eager/python/examples/rnn_ptb/ |
H A D | rnn_ptb.py | 66 input_seq = tf.nn.dropout(input_seq, self.keep_ratio) 116 num_layers, hidden_dim, dropout=dropout_ratio) 138 y = tf.nn.dropout(y, self.keep_ratio) 312 FLAGS.hidden_dim, FLAGS.num_layers, FLAGS.dropout, 353 "--dropout", type=float, default=0.2, help="Drop out ratio.")
|
/external/tensorflow/tensorflow/examples/learn/ |
H A D | iris_custom_model.py | 30 """DNN with three hidden layers, and dropout of 0.1 probability.""" 32 # each layer having a dropout probability of 0.1. 36 net = tf.layers.dropout(net, rate=0.1)
|
H A D | multiple_gpu.py | 34 """DNN with three hidden layers, and dropout of 0.1 probability. 48 # each layer having a dropout probability of 0.1. 53 net = tf.layers.dropout(net, rate=0.1)
|
/external/tensorflow/tensorflow/examples/tutorials/mnist/ |
H A D | mnist_with_summaries.py | 103 with tf.name_scope('dropout'): 106 dropped = tf.nn.dropout(hidden1, keep_prob) 153 k = FLAGS.dropout 198 parser.add_argument('--dropout', type=float, default=0.9, 199 help='Keep probability for training dropout.')
|
/external/tensorflow/tensorflow/python/layers/ |
H A D | core.py | 265 rate: The dropout rate, between 0 and 1. E.g. `rate=0.1` would drop out 268 binary dropout mask that will be multiplied with the input. 270 `(batch_size, timesteps, features)`, and you want the dropout mask 300 return nn.dropout(inputs, 1 - self.rate, 311 @tf_export('layers.dropout') 312 def dropout(inputs, function 327 rate: The dropout rate, between 0 and 1. E.g. "rate=0.1" would drop out 330 binary dropout mask that will be multiplied with the input. 332 `(batch_size, timesteps, features)`, and you want the dropout mask 340 (apply dropout) o [all...] |
/external/tensorflow/tensorflow/contrib/boosted_trees/python/kernel_tests/ |
H A D | prediction_ops_test.py | 188 # Empty dropout. 216 # Empty dropout. 246 # Empty dropout. 295 # Empty dropout. 383 # Empty dropout. 430 # Empty dropout. 479 # Empty dropout. 528 # Empty dropout. 576 # Empty dropout. 627 # Empty dropout [all...] |
/external/tensorflow/tensorflow/contrib/receptive_field/python/util/ |
H A D | receptive_field_test.py | 165 """Aligned network with dropout for test. 168 has dropout normalization. 184 dropout = slim.dropout(l3) 186 nn.relu(l1 + dropout, name='output') 383 # uniform variable of the dropout. 388 ['Dropout/dropout/random_uniform']))
|