| /external/tensorflow/tensorflow/python/keras/ |
| D | model_subclassing_test.py | 49 def __init__(self, use_bn=False, use_dp=False, num_classes=10): argument
53 self.num_classes = num_classes
56 self.dense2 = keras.layers.Dense(num_classes, activation='softmax')
73 def __init__(self, num_classes=10): argument
75 self.num_classes = num_classes
79 self.dense1 = keras.layers.Dense(num_classes, activation='softmax')
89 def __init__(self, use_bn=False, use_dp=False, num_classes=(2, 3)): argument
93 self.num_classes = num_classes
96 self.dense2 = keras.layers.Dense(num_classes[0], activation='softmax')
97 self.dense3 = keras.layers.Dense(num_classes[1], activation='softmax')
[all …]
|
| D | testing_utils.py | 43 num_classes, argument
51 num_classes: Integer, number of classes for the data and targets.
60 templates = 2 * num_classes * np.random.random((num_classes,) + input_shape)
61 y = np.random.randint(0, num_classes, size=(num_sample,))
288 def get_small_sequential_mlp(num_hidden, num_classes, input_dim=None): argument
295 activation = 'sigmoid' if num_classes == 1 else 'softmax'
296 model.add(keras.layers.Dense(num_classes, activation=activation))
300 def get_small_functional_mlp(num_hidden, num_classes, input_dim): argument
303 activation = 'sigmoid' if num_classes == 1 else 'softmax'
304 outputs = keras.layers.Dense(num_classes, activation=activation)(outputs)
[all …]
|
| D | callbacks_v1_test.py | 38 NUM_CLASSES = 2 variable
57 num_classes=NUM_CLASSES)
85 model.add(keras.layers.Dense(NUM_CLASSES, activation='softmax'))
166 num_classes=NUM_CLASSES)
192 output1 = keras.layers.Dense(NUM_CLASSES, activation='softmax')(hidden)
193 output2 = keras.layers.Dense(NUM_CLASSES, activation='softmax')(hidden)
270 num_classes=NUM_CLASSES)
281 model.add(keras.layers.Dense(NUM_CLASSES, activation='softmax'))
321 num_hidden=10, num_classes=10, input_dim=100)
366 num_classes=NUM_CLASSES)
[all …]
|
| /external/tensorflow/tensorflow/contrib/slim/python/slim/nets/ |
| D | inception_v3_test.py | 41 num_classes = 1000
44 logits, end_points = inception_v3.inception_v3(inputs, num_classes)
47 [batch_size, num_classes])
50 [batch_size, num_classes])
135 num_classes = 1000
138 _, end_points = inception_v3.inception_v3(inputs, num_classes)
142 [batch_size, num_classes])
146 [batch_size, num_classes])
159 num_classes = 1000
162 _, end_points = inception_v3.inception_v3(inputs, num_classes)
[all …]
|
| D | vgg_test.py | 36 num_classes = 1000
39 logits, _ = vgg.vgg_a(inputs, num_classes)
42 [batch_size, num_classes])
47 num_classes = 1000
50 logits, _ = vgg.vgg_a(inputs, num_classes, spatial_squeeze=False)
53 [batch_size, 2, 2, num_classes])
58 num_classes = 1000
62 _, end_points = vgg.vgg_a(inputs, num_classes, is_training=is_training)
75 num_classes = 1000
78 vgg.vgg_a(inputs, num_classes)
[all …]
|
| D | inception_v2_test.py | 41 num_classes = 1000
44 logits, end_points = inception_v2.inception_v2(inputs, num_classes)
47 [batch_size, num_classes])
50 [batch_size, num_classes])
129 num_classes = 1000
132 _, end_points = inception_v2.inception_v2(inputs, num_classes)
140 inputs, num_classes, scope='depth_multiplied_net', depth_multiplier=0.5)
150 num_classes = 1000
153 _, end_points = inception_v2.inception_v2(inputs, num_classes)
161 inputs, num_classes, scope='depth_multiplied_net', depth_multiplier=2.0)
[all …]
|
| D | overfeat_test.py | 35 num_classes = 1000
38 logits, _ = overfeat.overfeat(inputs, num_classes)
41 [batch_size, num_classes])
46 num_classes = 1000
49 logits, _ = overfeat.overfeat(inputs, num_classes, spatial_squeeze=False)
52 [batch_size, 2, 2, num_classes])
57 num_classes = 1000
60 _, end_points = overfeat.overfeat(inputs, num_classes)
72 num_classes = 1000
75 overfeat.overfeat(inputs, num_classes)
[all …]
|
| D | alexnet_test.py | 35 num_classes = 1000
38 logits, _ = alexnet.alexnet_v2(inputs, num_classes)
41 [batch_size, num_classes])
46 num_classes = 1000
49 logits, _ = alexnet.alexnet_v2(inputs, num_classes, spatial_squeeze=False)
52 [batch_size, 4, 7, num_classes])
57 num_classes = 1000
60 _, end_points = alexnet.alexnet_v2(inputs, num_classes)
72 num_classes = 1000
75 alexnet.alexnet_v2(inputs, num_classes)
[all …]
|
| D | inception_v1_test.py | 41 num_classes = 1000
44 logits, end_points = inception_v1.inception_v1(inputs, num_classes)
47 [batch_size, num_classes])
50 [batch_size, num_classes])
144 num_classes = 1000
149 logits, end_points = inception_v1.inception_v1(inputs, num_classes)
152 [batch_size, num_classes])
162 num_classes = 1000
165 logits, _ = inception_v1.inception_v1(inputs, num_classes)
167 self.assertListEqual(logits.get_shape().as_list(), [None, num_classes])
[all …]
|
| /external/google-fruit/extras/benchmark/tables/ |
| D | fruit_wiki.yml | 10 dimension: "num_classes"
92 num_classes: 100
111 num_classes: 1000
130 num_classes: 100
149 num_classes: 1000
168 num_classes: 100
187 num_classes: 1000
206 num_classes: 100
225 num_classes: 1000
244 num_classes: 100
[all …]
|
| /external/tensorflow/tensorflow/contrib/tensor_forest/kernels/v4/ |
| D | stat_utils.cc | 24 // num_classes for smoothing each class, then Gini looks more like this:
33 float GiniImpurity(const LeafStat& stats, int32 num_classes) { in GiniImpurity() argument
34 const float smoothed_sum = num_classes + stats.weight_sum(); in GiniImpurity()
36 2 * stats.weight_sum() + num_classes) / in GiniImpurity()
40 float WeightedGiniImpurity(const LeafStat& stats, int32 num_classes) { in WeightedGiniImpurity() argument
41 return stats.weight_sum() * GiniImpurity(stats, num_classes); in WeightedGiniImpurity()
74 float SmoothedGini(float sum, float square, int num_classes) { in SmoothedGini() argument
76 const float smoothed_sum = num_classes + sum; in SmoothedGini()
77 return 1.0 - (square + 2 * sum + num_classes) / (smoothed_sum * smoothed_sum); in SmoothedGini()
80 float WeightedSmoothedGini(float sum, float square, int num_classes) { in WeightedSmoothedGini() argument
[all …]
|
| /external/tensorflow/tensorflow/python/ops/ |
| D | confusion_matrix.py | 96 num_classes=None, argument
108 If `num_classes` is `None`, then `num_classes` will be set to one plus the
110 start at 0. For example, if `num_classes` is 3, then the possible labels
133 num_classes: The possible number of labels the classification task can
151 (predictions, labels, num_classes, weights)) as name:
169 if num_classes is None:
170 num_classes = math_ops.maximum(math_ops.reduce_max(predictions),
173 num_classes_int64 = math_ops.cast(num_classes, dtypes.int64)
189 shape = array_ops.stack([num_classes, num_classes])
206 num_classes=None, argument
[all …]
|
| D | nn_test.py | 507 def _GenerateTestData(self, num_classes, dim, batch_size, num_true, labels, argument
514 num_classes: An int. The number of embedding classes in the test case.
519 sampled: A list of indices in [0, num_classes).
525 of shape [num_classes, dim]
527 of shape [num_classes].
537 weights = np.random.randn(num_classes, dim).astype(np.float32)
538 biases = np.random.randn(num_classes).astype(np.float32)
593 num_classes = 5
598 low=0, high=num_classes, size=batch_size * num_true)
601 num_classes=num_classes,
[all …]
|
| /external/tensorflow/tensorflow/contrib/nn/python/ops/ |
| D | sampling_ops.py | 116 num_classes, argument
175 weights: A `Tensor` or `PartitionedVariable` of shape `[num_classes, dim]`,
177 has shape [num_classes, dim]. The (possibly-sharded) class embeddings.
178 biases: A `Tensor` or `PartitionedVariable` of shape `[num_classes]`.
189 num_classes: An `int`. The number of possible classes.
209 if num_sampled > num_classes:
210 raise ValueError("num_sampled ({}) cannot be greater than num_classes ({})".
211 format(num_sampled, num_classes))
227 range_max=num_classes)
239 num_classes=num_classes,
[all …]
|
| /external/tensorflow/tensorflow/python/keras/utils/ |
| D | np_utils_test.py | 30 num_classes = 5
32 expected_shapes = [(1, num_classes),
33 (3, num_classes),
34 (4, 3, num_classes),
35 (5, 4, 3, num_classes),
36 (3, num_classes)]
37 labels = [np.random.randint(0, num_classes, shape) for shape in shapes]
39 keras.utils.to_categorical(label, num_classes) for label in labels]
|
| D | np_utils.py | 25 def to_categorical(y, num_classes=None, dtype='float32'): argument
32 (integers from 0 to num_classes).
33 num_classes: total number of classes.
45 if not num_classes:
46 num_classes = np.max(y) + 1
48 categorical = np.zeros((n, num_classes), dtype=dtype)
50 output_shape = input_shape + (num_classes,)
|
| /external/tensorflow/tensorflow/core/util/ctc/ |
| D | ctc_loss_calculator.h | 94 int num_classes, const Vector& seq_len,
123 auto num_classes = inputs[0].cols(); in CalculateLoss() local
136 if (inputs[t].cols() != num_classes) { in CalculateLoss()
138 " to be: ", num_classes, in CalculateLoss()
161 batch_size, num_classes, seq_len, labels, &max_u_prime, &l_primes); in CalculateLoss()
167 auto ComputeLossAndGradients = [this, num_classes, &labels, &l_primes, in CalculateLoss()
196 Matrix y(num_classes, seq_len(b)); in CalculateLoss()
206 // y, prob are in num_classes x seq_len(b) in CalculateLoss()
262 max_seq_len * num_classes * in CalculateLoss()
264 max_seq_len * 2 * (2 * num_classes + 1) * in CalculateLoss()
[all …]
|
| /external/tensorflow/tensorflow/contrib/tensor_forest/kernels/ |
| D | tree_utils_test.cc | 98 const int32 num_classes = 4; in TEST() local
102 {num_accumulators, num_classes}); in TEST()
108 {num_accumulators, num_splits, num_classes}); in TEST()
116 const int32 num_classes = 4; in TEST() local
121 {num_accumulators, num_classes}); in TEST()
127 {num_accumulators, num_splits, num_classes}); in TEST()
135 const int32 num_classes = 4; in TEST() local
139 {num_accumulators, num_classes}); in TEST()
143 {num_accumulators, num_classes}); in TEST()
150 {num_accumulators, num_splits, num_classes}); in TEST()
[all …]
|
| D | tree_utils.cc | 68 const Eigen::Tensor<float, 1, Eigen::RowMajor>& rights, int32 num_classes, in ClassificationSplitScore() argument
72 // count vector is num_classes + 1. in ClassificationSplitScore()
73 offsets[0] = i * (num_classes + 1) + 1; in ClassificationSplitScore()
75 extents[0] = num_classes; in ClassificationSplitScore()
86 const int32 num_classes = in GetTwoBestClassification() local
106 std::bind(ClassificationSplitScore, splits, rights, num_classes, in GetTwoBestClassification()
238 const int32 num_classes = in MakeBootstrapWeights() local
246 float denom = static_cast<float>(n) + static_cast<float>(num_classes); in MakeBootstrapWeights()
248 weights->resize(num_classes * 2); in MakeBootstrapWeights()
249 for (int i = 0; i < num_classes; i++) { in MakeBootstrapWeights()
[all …]
|
| /external/tensorflow/tensorflow/core/kernels/ |
| D | multinomial_op_gpu.cu.cc | 42 __global__ void MultinomialKernel(int32 nthreads, const int32 num_classes, in MultinomialKernel() argument
46 const int maxima_idx = index / num_classes; in MultinomialKernel()
50 static_cast<UnsignedOutputType>(index % num_classes)); in MultinomialKernel()
62 int num_classes, int num_samples, in operator ()()
74 bsc.set(2, num_classes); in operator ()()
78 boc.set(2, num_classes); in operator ()()
83 Eigen::array<int, 3> bsc{batch_size, num_samples, num_classes}; in operator ()()
84 Eigen::array<int, 3> boc{batch_size, 1, num_classes}; in operator ()()
106 /*in_dim1=*/num_classes, /*in_dim2=*/1, /*out_rank=*/1, in operator ()()
112 const int32 work_items = batch_size * num_samples * num_classes; in operator ()()
[all …]
|
| D | multinomial_op.cc | 51 int num_classes, int num_samples,
77 int num_classes, int num_samples, in operator ()()
86 auto DoWork = [ctx, num_samples, num_classes, &gen, &output, &logits]( in operator ()()
99 ctx->allocate_temp(DT_DOUBLE, TensorShape({num_classes}), in operator ()()
107 for (int64 j = 0; j < num_classes; ++j) { in operator ()()
119 for (int64 j = 0; j < num_classes; ++j) { in operator ()()
127 const double* cdf_end = cdf.data() + num_classes; in operator ()()
137 50 * (num_samples * std::log(num_classes) / std::log(2) + num_classes); in operator ()()
176 const int num_classes = static_cast<int>(logits_t.dim_size(1)); in DoCompute() local
177 OP_REQUIRES(ctx, num_classes > 0, in DoCompute()
[all …]
|
| /external/tensorflow/tensorflow/python/kernel_tests/random/ |
| D | multinomial_op_test.py | 154 logits: Numpy ndarray of shape [batch_size, num_classes].
156 sampler: A sampler function that takes (1) a [batch_size, num_classes]
160 Frequencies from sampled classes; shape [batch_size, num_classes].
167 batch_size, num_classes = logits.shape
173 self.assertLess(max(cnts.keys()), num_classes)
177 for k in range(num_classes)]
203 with self.assertRaisesOpError("num_classes should be positive"):
216 def native_op_vs_composed_ops(batch_size, num_classes, num_samples, num_iters): argument
218 shape = [batch_size, num_classes]
247 for num_classes in [10000, 100000]:
[all …]
|
| /external/tensorflow/tensorflow/contrib/boosted_trees/estimator_batch/ |
| D | estimator_test.py | 146 learner_config.num_classes = 2
165 learner_config.num_classes = 2
187 learner_config.num_classes = 2
212 learner_config.num_classes = 2
232 learner_config.num_classes = 2
252 learner_config.num_classes = 2
280 learner_config.num_classes = 2
300 learner_config.num_classes = 2
320 learner_config.num_classes = 3
330 n_classes=learner_config.num_classes,
[all …]
|
| /external/google-fruit/extras/benchmark/suites/ |
| D | fruit_full.yml | 24 num_classes: &num_classes
48 num_classes: *num_classes
62 num_classes: *num_classes
|
| /external/tensorflow/tensorflow/python/keras/engine/ |
| D | sequential_test.py | 67 num_classes = 2
70 num_hidden, num_classes, input_dim)
76 y = np.random.random((batch_size, num_classes))
105 num_classes = 2
107 model = testing_utils.get_small_sequential_mlp(num_hidden, num_classes)
118 y = np.random.random((batch_size, num_classes))
128 num_classes = 2
132 model = testing_utils.get_small_sequential_mlp(num_hidden, num_classes)
143 y = array_ops.zeros((num_samples, num_classes))
227 num_classes = 2
[all …]
|