COMPMID-470: Neon Deconvolution.

Implemented by up-sampling the input with zeros insertions between the input samples and convolving the Deconvolution kernels on the up-sampled result.

The upsampling is performed by the function NEDeconvolutionLayerUpsample.
Convolving is done by NEDirectConvolutionLayer.

Change-Id: I25f7ba7c6b99cd9310797972ede40aeff4a54900
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/85319
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>

1 files changed, 96 insertions, 0 deletions

diff --git a/arm_compute/runtime/NEON/functions/NEDeconvolutionLayer.h b/arm_compute/runtime/NEON/functions/NEDeconvolutionLayer.h
new file mode 100644
index 0000000000..3433e77ba1
--- /dev/null
+++ b/arm_compute/runtime/NEON/functions/NEDeconvolutionLayer.h
@@ -0,0 +1,96 @@
+/*
+ * Copyright (c) 2017 ARM Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#ifndef __ARM_COMPUTE_NEDECONVOLUTIONLAYER_H__
+#define __ARM_COMPUTE_NEDECONVOLUTIONLAYER_H__
+
+#include "arm_compute/runtime/NEON/functions/NEDeconvolutionLayerUpsample.h"
+#include "arm_compute/runtime/NEON/functions/NEDirectConvolutionLayer.h"
+
+#include "arm_compute/core/Types.h"
+#include "arm_compute/runtime/IFunction.h"
+#include "arm_compute/runtime/IMemoryManager.h"
+#include "arm_compute/runtime/MemoryGroup.h"
+#include "arm_compute/runtime/Tensor.h"
+
+#include <memory>
+
+namespace arm_compute
+{
+/** Function to run the deconvolution layer.
+ *
+ *  The operation is similar to convolution but it's implemented by up-sampling the inputs with zeros insertions between the inputs and convolving
+ *  the kernels on the up-sampled result.
+ *
+ *  Before the Deconvolution is done, up-scaling the first 2D with zeros is performed. The relation between input to
+ *  output is as follows:
+ *      width_output = round((width_input − 1) ∗ upscale_x − 2 ∗ padding_x + kernel_x + a_x )
+ *      height_output = round((height_input − 1) ∗ upscale_y − 2 ∗ padding_y + kernel_y + a_y )
+ *
+ *  where
+ *      width is the size of the first input dimension.
+ *      height is the size of the second input dimension.
+ *      width_output is the size of the first output dimension.
+ *      height_output is the size of the second output dimension.
+ *      kernel_x and kernel_y are the convolution sizes in x and y.
+ *      ax and ay the number of zeros added to the top and right edges of the input.
+ *      upscale_x and upscale_y how much to scale the X and Y axis.
+ *
+ *  This function calls the following NEON kernels:
+ *
+ * -# @ref NEDeconvolutionLayerUpsampleKernel
+ * -# @ref NEDirectConvolutionLayer
+ *
+ */
+class NEDeconvolutionLayer : public IFunction
+{
+public:
+    /** Constructor */
+    NEDeconvolutionLayer(std::shared_ptr<IMemoryManager> memory_manager = nullptr);
+    /** Set the input, weights, biases and output tensors.
+      *
+      * @param[in,out] input    Input tensor. 3 lower dimensions represent a single input, and an optional 4th dimension for batch of inputs. Data types supported: F32.
+      * @param[in]     weights  The 4d weights with dimensions [width, height, OFM, IFM]. Data type supported: Same as @p input.
+      * @param[in]     bias     Optional, ignored if NULL. The biases have one dimension. Data type supported: Same as @p input.
+      * @param[out]    output   Output tensor. The output has the same number of dimensions as the @p input.
+      * @param[in]     info     Contains padding and policies to be used in the deconvolution, this is decribed in @ref PadStrideInfo.
+      * @param[in]     ax       The number of zeros added to right edge of the input.
+      * @param[in]     ay       The number of zeros added to top edge of the input.
+      * @param[in]     upscalex How much to scale the X axis.
+      * @param[in]     upscaley How much to scale the Y axis.
+      *
+      */
+    void configure(ITensor *input, const ITensor *weights, const ITensor *bias, ITensor *output, const PadStrideInfo &info,
+                   unsigned int ax, unsigned int ay, float upscalex, float upscaley);
+
+    // Inherited methods overridden:
+    void run() override;
+
+private:
+    MemoryGroup                  _memory_group;
+    NEDeconvolutionLayerUpsample _scale_f;
+    NEDirectConvolutionLayer     _conv_f;
+    Tensor                       _scaled_output;
+};
+} // arm_compute
+#endif /* __ARM_COMPUTE_NEDECONVOLUTIONLAYER_H__ */