COMPMID-1364: Add support for NHWC in NEDepthConcatenateLayer
Change-Id: I4f8e46d1c79afa9284f2c6dc00383c453a8e7bd5
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/140165
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-by: Pablo Tello <pablo.tello@arm.com>
Tested-by: Jenkins <bsgcomp@arm.com>
diff --git a/arm_compute/core/NEON/NEKernels.h b/arm_compute/core/NEON/NEKernels.h
index b9c54b2..156b116 100644
--- a/arm_compute/core/NEON/NEKernels.h
+++ b/arm_compute/core/NEON/NEKernels.h
@@ -112,6 +112,7 @@
#include "arm_compute/core/NEON/kernels/NETransposeKernel.h"
#include "arm_compute/core/NEON/kernels/NEWarpKernel.h"
#include "arm_compute/core/NEON/kernels/NEWeightsReshapeKernel.h"
+#include "arm_compute/core/NEON/kernels/NEWidthConcatenateLayerKernel.h"
#include "arm_compute/core/NEON/kernels/NEWinogradConvolutionLayerKernel.h"
#endif /* __ARM_COMPUTE_NEKERNELS_H__ */
diff --git a/arm_compute/core/NEON/kernels/NEDepthConcatenateLayerKernel.h b/arm_compute/core/NEON/kernels/NEDepthConcatenateLayerKernel.h
index 12a5051..848d89f 100644
--- a/arm_compute/core/NEON/kernels/NEDepthConcatenateLayerKernel.h
+++ b/arm_compute/core/NEON/kernels/NEDepthConcatenateLayerKernel.h
@@ -55,7 +55,7 @@
~NEDepthConcatenateLayerKernel() = default;
/** Initialise the kernel's inputs and output
*
- * @param[in] input Input tensor. Data types supported: F16/F32.
+ * @param[in] input Input tensor. Data types supported: QASYMM8/F16/F32.
* @param[in] depth_offset The offset on the Z axis.
* @param[in,out] output Output tensor. Data types supported: Same as @p input.
*
@@ -64,6 +64,15 @@
*
*/
void configure(const ITensor *input, unsigned int depth_offset, ITensor *output);
+ /** Static function to check if given info will lead to a valid configuration of @ref NEDepthConcatenateLayerKernel
+ *
+ * @param[in] input Input tensor info. Data types supported: QASYMM8/F16/F32.
+ * @param[in] depth_offset The offset on the Z axis.
+ * @param[in] output Output tensor info. Data types supported: Same as @p input.
+ *
+ * @return a status
+ */
+ static Status validate(const ITensorInfo *input, unsigned int depth_offset, const ITensorInfo *output);
// Inherited methods overridden:
void run(const Window &window, const ThreadInfo &info) override;
diff --git a/arm_compute/core/NEON/kernels/NEWidthConcatenateLayerKernel.h b/arm_compute/core/NEON/kernels/NEWidthConcatenateLayerKernel.h
new file mode 100644
index 0000000..4cf3273
--- /dev/null
+++ b/arm_compute/core/NEON/kernels/NEWidthConcatenateLayerKernel.h
@@ -0,0 +1,84 @@
+/*
+ * Copyright (c) 2018 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_NEWIDTHCONCATENATELAYERKERNEL_H__
+#define __ARM_COMPUTE_NEWIDTHCONCATENATELAYERKERNEL_H__
+
+#include "arm_compute/core/NEON/INEKernel.h"
+#include "arm_compute/core/Types.h"
+
+namespace arm_compute
+{
+class ITensor;
+
+/** Interface for the width concatenate kernel.
+ * The input tensor will be concatenated into the output tensor.
+ */
+class NEWidthConcatenateLayerKernel : public INEKernel
+{
+public:
+ const char *name() const override
+ {
+ return "NEWidthConcatenateLayerKernel";
+ }
+ /** Default constructor */
+ NEWidthConcatenateLayerKernel();
+ /** Prevent instances of this class from being copied (As this class contains pointers) */
+ NEWidthConcatenateLayerKernel(const NEWidthConcatenateLayerKernel &) = delete;
+ /** Prevent instances of this class from being copied (As this class contains pointers) */
+ NEWidthConcatenateLayerKernel &operator=(const NEWidthConcatenateLayerKernel &) = delete;
+ /** Allow instances of this class to be moved */
+ NEWidthConcatenateLayerKernel(NEWidthConcatenateLayerKernel &&) = default;
+ /** Allow instances of this class to be moved */
+ NEWidthConcatenateLayerKernel &operator=(NEWidthConcatenateLayerKernel &&) = default;
+ /** Default destructor */
+ ~NEWidthConcatenateLayerKernel() = default;
+ /** Initialise the kernel's inputs and output
+ *
+ * @param[in] input Input tensor. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32
+ * @param[in] width_offset The offset on the X axis.
+ * @param[in,out] output Output tensor. Data types supported: Same as @p input.
+ *
+ */
+ void configure(const ITensor *input, unsigned int width_offset, ITensor *output);
+ /** Static function to check if given info will lead to a valid configuration of @ref NEWidthConcatenateLayerKernel
+ *
+ * @param[in] input Input tensor info. Data types supported: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32
+ * @param[in] width_offset The offset on the X axis.
+ * @param[in] output Output tensor info. Data types supported: Same as @p input.
+ *
+ * @return a status
+ */
+ static Status validate(const ITensorInfo *input, unsigned int width_offset, const ITensorInfo *output);
+
+ // Inherited methods overridden:
+ void run(const Window &window, const ThreadInfo &info) override;
+
+private:
+ const ITensor *_input;
+ ITensor *_output;
+ unsigned int _width_offset;
+};
+} // namespace arm_compute
+#endif /* __ARM_COMPUTE_NEWIDTHCONCATENATELAYERKERNEL_H__ */