| /external/tensorflow/tensorflow/compiler/xla/service/ |
| hlo_matchers.h | 112 HLO_MATCHER(Convolution);
|
| /external/tensorflow/tensorflow/contrib/kfac/python/ops/ |
| fisher_blocks.py | 473 # Infer number of locations upon which convolution is applied.
507 the convolution.
783 # Infer number of locations upon which convolution is applied.
810 the convolution.
[all...] |
| /external/tensorflow/tensorflow/core/grappler/costs/ |
| utils_test.cc | 105 // Now add the convolution op
|
| /external/tensorflow/tensorflow/core/kernels/ |
| resize_bicubic_op_test.cc | 62 // convolution algorithm.
|
| xsmm_conv2d_test.cc | 240 // Run naive convolution
|
| eigen_backward_cuboid_convolutions.h | 27 * \brief Computes the backprop for the input of a 3D convolution.
305 * \brief Computes the backprop for the filter of a 3D convolution.
|
| /external/tensorflow/tensorflow/docs_src/extend/tool_developers/ |
| index.md | 179 The ordering of convolution weight values is often tricky to deal with when
|
| /external/tensorflow/tensorflow/docs_src/tutorials/ |
| layers.md | 49 * **Convolutional layers**, which apply a specified number of convolution
272 a 5x5 convolution over a 28x28 tensor will produce a 24x24 tensor, as there are
276 output of the convolution. Here, we specify ReLU activation with
353 convolution/pooling layers. Before we connect the layer, however, we'll flatten
|
| /external/tensorflow/tensorflow/python/keras/_impl/keras/layers/ |
| advanced_activations.py | 95 are from a 2D convolution
|
| /external/tensorflow/tensorflow/python/keras/_impl/keras/utils/ |
| layer_utils.py | 207 """Converts all convolution kernels in a model from Theano to TensorFlow.
|
| /external/tensorflow/tensorflow/python/kernel_tests/ |
| atrous_conv2d_test.py | 15 """Tests for convolution related functionality in tensorflow.ops.nn."""
|
| conv3d_transpose_test.py | 15 """Tests for convolution related functionality in tensorflow.ops.nn."""
|
| /external/tensorflow/tensorflow/tools/graph_transforms/ |
| fold_old_batch_norms.cc | 147 // The input and convolution can be copied straight over, since the
|
| /frameworks/rs/cpu_ref/ |
| rsCpuIntrinsics_advsimd_Blur.S | 54 * The convolution is fully unrolled for windows up to max_r, with the
56 * into the relevant part of the code for an arbitrary convolution radius. Two
236 /* Some portion of the convolution window (as much as will fit, and all of it
254 * x9 -- pointer to additional convolution window data
[all...] |
| rsCpuIntrinsics_neon_Blur.S | 58 * The convolution is fully unrolled for windows up to max_r, with the
60 * into the relevant part of the code for an arbitrary convolution radius. Two
242 /* Some portion of the convolution window (as much as will fit, and all of it
260 * r9 -- pointer to additional convolution window data
[all...] |
| /external/ImageMagick/ImageMagick/api/ |
| effect.html | 164 <p>ConvolveImage() applies a custom convolution kernel to the image.</p>
221 <p>EdgeImage() finds edges in an image. Radius defines the radius of the convolution filter. Use a radius of 0 and EdgeImage() selects a suitable radius for you.</p>
|
| /external/ImageMagick/www/api/ |
| effect.php | 164 <p>ConvolveImage() applies a custom convolution kernel to the image.</p>
221 <p>EdgeImage() finds edges in an image. Radius defines the radius of the convolution filter. Use a radius of 0 and EdgeImage() selects a suitable radius for you.</p>
|
| /external/libopus/silk/fixed/ |
| pitch_analysis_core_FIX.c | 287 /* Convolution */
301 /* Convolution */
[all...] |
| /external/libopus/silk/float/ |
| pitch_analysis_core_FLP.c | 250 /* Convolution */
263 /* Convolution */
|
| /external/opencv/cv/src/ |
| cvoptflowlk.cpp | 128 /* buffers for gaussian horisontal convolution */
369 /* ---------Calculating horizontal convolution of processed line----------------------- */
|
| /external/tensorflow/tensorflow/contrib/fused_conv/kernels/ |
| fused_conv2d_bias_activation_op.cc | 252 // A dummy type to group forward convolution autotune results together.
291 // convolution. I don't know cuDNN7 still has the bug, so enable this
|
| /external/tensorflow/tensorflow/core/ops/ |
| nn_ops.cc | 372 // Work out the convolution's effect with 'padded' as the input.
[all...] |
| /external/tensorflow/tensorflow/examples/android/jni/object_tracking/ |
| image-inl.h | 492 // Convolution loop unrolled for performance...
510 // Convolution loop unrolled for performance...
|
| /external/tensorflow/tensorflow/python/framework/ |
| common_shapes.py | 107 """Returns the spatial size of a n-d convolution/pooling output."""
150 """Returns the number of rows and columns in a convolution/pooling output."""
|
| /external/tensorflow/tensorflow/stream_executor/cuda/ |
| cuda_dnn.h | 370 LOG(ERROR) << "separable convolution not supported by CUDNN";
575 // Derives an output batch descriptor from an input batch and convolution
[all...] |
