/external/tensorflow/tensorflow/compiler/tf2xla/kernels/ |
D | extract_image_patches_op.cc | 78 int input_dim = GetTensorSpatialDimIndex(num_dims, data_format, i); in Compile() local 80 ctx, ksizes_[input_dim] >= 0, in Compile() 83 dilations_[input_dim])); in Compile() 84 OP_REQUIRES(ctx, strides_[input_dim] >= 1, in Compile() 87 dilations_[input_dim])); in Compile() 88 OP_REQUIRES(ctx, dilations_[input_dim] >= 1, in Compile() 91 dilations_[input_dim])); in Compile() 111 int input_dim = GetTensorSpatialDimIndex(num_dims, data_format, i); in Compile() local 112 kernel_shape[i] = ksizes_[input_dim]; in Compile() 113 kernel_size *= ksizes_[input_dim]; in Compile()
|
D | conv_op_helpers.cc | 124 int input_dim = GetTensorSpatialDimIndex(num_dims, attrs.data_format, i); in CheckConvAttrs() local 125 if (attrs.dilations[input_dim] < 1) { in CheckConvAttrs() 128 attrs.dilations[input_dim]); in CheckConvAttrs()
|
/external/tensorflow/tensorflow/python/keras/layers/ |
D | embeddings.py | 107 input_dim, argument 121 if input_dim <= 0 or output_dim <= 0: 124 input_dim, output_dim)) 137 self.input_dim = input_dim 150 shape=(self.input_dim, self.output_dim), 200 'input_dim': self.input_dim,
|
D | convolutional.py | 976 input_dim = int(input_shape[channel_axis]) 977 self.input_spec = InputSpec(ndim=3, axes={channel_axis: input_dim}) 978 kernel_shape = self.kernel_size + (self.filters, input_dim) 1247 input_dim = int(input_shape[channel_axis]) 1248 self.input_spec = InputSpec(ndim=4, axes={channel_axis: input_dim}) 1249 kernel_shape = self.kernel_size + (self.filters, input_dim) 1557 input_dim = int(input_shape[channel_axis]) 1558 kernel_shape = self.kernel_size + (self.filters, input_dim) 1559 self.input_spec = InputSpec(ndim=5, axes={channel_axis: input_dim}) 1821 input_dim = int(input_shape[channel_axis]) [all …]
|
D | cudnn_recurrent.py | 243 input_dim = int(input_shape[-1]) 246 shape=(input_dim, self.units * 3), 429 input_dim = int(input_shape[-1]) 432 shape=(input_dim, self.units * 4),
|
D | convolutional_recurrent.py | 549 input_dim = input_shape[channel_axis] 550 kernel_shape = self.kernel_size + (input_dim, self.filters * 4)
|
D | recurrent.py | 1799 input_dim = input_shape[-1] 1802 shape=(input_dim, self.units * 3), 2362 input_dim = input_shape[-1] 2364 shape=(input_dim, self.units * 4),
|
/external/eigen/unsupported/Eigen/CXX11/src/Tensor/ |
D | TensorConvolutionSycl.h | 328 const Index input_dim = input_dims[index]; 330 const Index result_dim = input_dim - kernel_dim + 1; 394 const auto input_dim = std::array<size_t, 2>{numX, numP}; 399 m_device.parallel_for_setup(input_dim, global_range, local_range); 412 indexMapper, kernel_size, cl::sycl::range<2>(input_dim[0], input_dim[1])); 424 auto input_dim = std::array<size_t, 3>{numX, numY, numP}; 429 m_device.parallel_for_setup(input_dim, global_range, local_range); 443 indexMapper, kernel_size, cl::sycl::range<3>{input_dim[0], input_dim[1], input_dim[2]}); 459 auto input_dim = std::array<size_t, 3>{numX, numY, numZ}; 473 m_device.parallel_for_setup(input_dim, global_range, local_range); [all …]
|
D | TensorDeviceSycl.h | 546 const std::array<Index, 2> &input_dim, cl::sycl::range<2> &global_range, in parallel_for_setup() argument 548 std::array<Index, 2> input_range = input_dim; in parallel_for_setup() 558 input_range[1] = input_dim[1]; in parallel_for_setup() 569 input_range[0] = input_dim[0]; in parallel_for_setup() 585 const std::array<Index, 3> &input_dim, cl::sycl::range<3> &global_range, in parallel_for_setup() argument 587 std::array<Index, 3> input_range = input_dim; in parallel_for_setup() 597 input_range[2] = input_dim[2]; in parallel_for_setup() 611 input_range[1] = input_dim[1]; in parallel_for_setup() 623 input_range[0] = input_dim[0]; in parallel_for_setup() 861 const std::array<Index, 2> &input_dim, cl::sycl::range<2> &global_range, in parallel_for_setup() [all …]
|
D | TensorConvolution.h | 33 const Index input_dim = input_dims[index]; in IndexMapper() local 35 const Index result_dim = input_dim - kernel_dim + 1; in IndexMapper() 347 const Index input_dim = input_dims[index]; 349 const Index result_dim = input_dim - kernel_dim + 1; 366 const Index input_dim = input_dims[index]; 368 const Index result_dim = input_dim - kernel_dim + 1; 811 const Index input_dim = input_dims[index]; 813 const Index result_dim = input_dim - kernel_dim + 1;
|
/external/XNNPACK/src/operators/ |
D | constant-pad-nd.c | 147 const size_t input_dim = input_shape[num_dims - 1 - i]; in setup_constant_pad_nd() local 152 normalized_input_shape[XNN_MAX_TENSOR_DIMS - 1 - num_squeezed_dims] = input_dim; in setup_constant_pad_nd() 153 …output_shape[XNN_MAX_TENSOR_DIMS - 1 - num_squeezed_dims] = pre_padding + input_dim + post_padding; in setup_constant_pad_nd() 163 normalized_input_shape[XNN_MAX_TENSOR_DIMS - num_squeezed_dims] *= input_dim; in setup_constant_pad_nd() 164 normalized_output_shape[XNN_MAX_TENSOR_DIMS - num_squeezed_dims] *= input_dim; in setup_constant_pad_nd()
|
/external/tensorflow/tensorflow/lite/kernels/internal/ |
D | batch_to_space_nd_test.cc | 25 int input_dim, int output_dim) { in GetIndexRange() argument 28 optimized_ops::GetIndexRange(spatial_index_dim, block_shape_dim, input_dim, in GetIndexRange()
|
/external/tensorflow/tensorflow/compiler/xla/service/ |
D | dynamic_padder.cc | 316 HloInstruction* reshape, int64_t input_dim, in GenerateBinaryMask() argument 325 ShapeUtil::MakeShape(xla::S32, {input_shape.dimensions(input_dim)}); in GenerateBinaryMask() 327 ShapeUtil::MakeShape(xla::PRED, {input_shape.dimensions(input_dim)}); in GenerateBinaryMask() 469 HloInstruction* reshape, int64_t input_dim, in RewriteDynamicReshapeSplitInput() argument 473 VLOG(2) << "Reshaping input dim " << input_dim << " to " in RewriteDynamicReshapeSplitInput() 479 ShapeUtil::MakeShape(xla::S32, {operand_shape.dimensions(input_dim)}); in RewriteDynamicReshapeSplitInput() 481 ShapeUtil::MakeShape(xla::PRED, {operand_shape.dimensions(input_dim)}); in RewriteDynamicReshapeSplitInput() 490 GenerateBinaryMask(reshape, input_dim, output_dims, output_dynamic_dims, in RewriteDynamicReshapeSplitInput() 515 dim->set_size(operand_shape.dimensions(input_dim)); in RewriteDynamicReshapeSplitInput() 517 dim->set_padding_low(operand_shape.dimensions(input_dim) - 1); in RewriteDynamicReshapeSplitInput() [all …]
|
D | shape_inference.cc | 3115 for (int64_t input_dim = 0; input_dim < operand.rank(); ++input_dim) { in InferReshapeShape() local 3116 if (!operand.is_dynamic_dimension(input_dim)) { in InferReshapeShape() 3124 input_dim); in InferReshapeShape() 3134 if (input_dim >= start.first && input_dim < end.first) { in InferReshapeShape() 3186 if (operand.dimensions(input_dim) == 1 && !new_sizes.empty()) { in InferReshapeShape() 3189 if (input_dim == 0) { in InferReshapeShape() 3192 if (input_dim == operand.rank() - 1) { in InferReshapeShape()
|
/external/tensorflow/tensorflow/lite/delegates/xnnpack/ |
D | reshape_tester.h | 35 for (int32_t input_dim : input_shape) { in InputShape() local 36 EXPECT_GT(input_dim, 0); in InputShape()
|
/external/tensorflow/tensorflow/python/keras/ |
D | testing_utils.py | 422 def get_small_sequential_mlp(num_hidden, num_classes, input_dim=None): argument 424 if input_dim: 425 model.add(layers.Dense(num_hidden, activation='relu', input_dim=input_dim)) 433 def get_small_functional_mlp(num_hidden, num_classes, input_dim): argument 434 inputs = layers.Input(shape=(input_dim,)) 499 def get_small_mlp(num_hidden, num_classes, input_dim): argument 507 return get_small_sequential_mlp(num_hidden, num_classes, input_dim) 509 return get_small_functional_mlp(num_hidden, num_classes, input_dim)
|
/external/tensorflow/tensorflow/python/kernel_tests/array_ops/ |
D | broadcast_to_ops_test.py | 53 for input_dim in range(1, 6): 54 for output_dim in range(input_dim, 6): 56 input_shape = [2] * input_dim
|
/external/XNNPACK/test/ |
D | constant-pad-operator-tester.h | 34 inline size_t input_dim(size_t i) const { in input_dim() function 84 return pre_padding(i) + input_dim(i) + post_padding(i); in output_dim() 123 input_dims[XNN_MAX_TENSOR_DIMS - num_dims() + i] = input_dim(i); in TestX8() 238 input_dims[XNN_MAX_TENSOR_DIMS - num_dims() + i] = input_dim(i); in TestX16() 353 input_dims[XNN_MAX_TENSOR_DIMS - num_dims() + i] = input_dim(i); in TestX32()
|
/external/tensorflow/tensorflow/python/debug/examples/v2/ |
D | debug_mnist_v2.py | 160 def get_dense_weights(input_dim, output_dim): argument 164 kernel = tf.Variable(initial_kernel([input_dim, output_dim]))
|
/external/tensorflow/tensorflow/python/debug/examples/v1/ |
D | debug_mnist_v1.py | 158 def nn_layer(input_tensor, input_dim, output_dim, layer_name, act=tf.nn.relu): argument 164 weights = weight_variable([input_dim, output_dim])
|
/external/tensorflow/tensorflow/core/kernels/linalg/ |
D | einsum_op_impl.h | 83 const int64_t input_dim = input.dim_size(axis); in RecordLabelToDimension() local 86 label_to_dim_sizes->at(label) != input_dim) { in RecordLabelToDimension() 90 " but got dimension ", input_dim); in RecordLabelToDimension() 92 (*label_to_dim_sizes)[label] = input_dim; in RecordLabelToDimension()
|
/external/tensorflow/tensorflow/compiler/xla/ |
D | shape_util.cc | 1471 for (int64_t input_dim = 0; input_dim < input_shape.rank(); ++input_dim) { in ReshapeIsBitcast() local 1472 if (input_shape.dimensions(input_dim) <= 1) { in ReshapeIsBitcast() 1477 input_unit_index[input_dim] = 1; in ReshapeIsBitcast() 1567 const int64_t input_dim = input_shape.layout().minor_to_major(input_minor); in AlignLayouts() local 1568 const int64_t common_factor = input_to_factor[input_dim]; in AlignLayouts()
|
/external/tensorflow/tensorflow/python/ops/signal/ |
D | fft_ops.py | 86 paddings = [[0, max(fft_dim.value - input_dim.value, 0)] 87 for fft_dim, input_dim in zip(
|
/external/tensorflow/tensorflow/python/kernel_tests/nn_ops/ |
D | depthwise_conv_op_base.py | 112 def PaddingsForDim(input_dim, filter_dim, stride): argument 114 if input_dim % stride == 0: 117 total_padding = max(filter_dim - (input_dim % stride), 0)
|
/external/tensorflow/tensorflow/compiler/mlir/tensorflow/ir/ |
D | tf_ops_a_m.cc | 411 int64_t input_dim = input_shape[spatial_dim_index]; in verify() local 413 if (!static_dims(input_dim, output_dim)) return success(); in verify() 415 int64_t input_dim_pad = input_dim * block_size; in verify() 418 if (crops_values.empty() && output_dim > input_dim * block_size) in verify() 424 << output_dim << ", input " << dim_name << " " << input_dim in verify() 438 << " " << input_dim << ", " << crop_a_name << " " << crop_a in verify()
|