COMPMID-1539 Implement YOLOLayer on CL
Change-Id: I332c0703e1399fca0c5b724529b54a28f49c88da
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/146842
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
diff --git a/arm_compute/core/CL/CLHelpers.h b/arm_compute/core/CL/CLHelpers.h
index 18d6bdf..a86870a 100644
--- a/arm_compute/core/CL/CLHelpers.h
+++ b/arm_compute/core/CL/CLHelpers.h
@@ -47,6 +47,14 @@
*/
std::string get_cl_type_from_data_type(const DataType &dt);
+/** Translates a tensor data type to the appropriate OpenCL select type.
+ *
+ * @param[in] dt @ref DataType to be translated to OpenCL select type.
+ *
+ * @return The string specifying the OpenCL select type to be used.
+ */
+std::string get_cl_select_type_from_data_type(const DataType &dt);
+
/** Get the size of a data type in number of bits.
*
* @param[in] dt @ref DataType.
diff --git a/arm_compute/core/CL/CLKernels.h b/arm_compute/core/CL/CLKernels.h
index 4a6773a..9586d8c 100644
--- a/arm_compute/core/CL/CLKernels.h
+++ b/arm_compute/core/CL/CLKernels.h
@@ -123,6 +123,7 @@
#include "arm_compute/core/CL/kernels/CLWinogradFilterTransformKernel.h"
#include "arm_compute/core/CL/kernels/CLWinogradInputTransformKernel.h"
#include "arm_compute/core/CL/kernels/CLWinogradOutputTransformKernel.h"
+#include "arm_compute/core/CL/kernels/CLYOLOLayerKernel.h"
#include "arm_compute/core/CL/kernels/ICLDepthwiseConvolutionLayer3x3Kernel.h"
#endif /* __ARM_COMPUTE_CLKERNELS_H__ */
diff --git a/arm_compute/core/CL/kernels/CLYOLOLayerKernel.h b/arm_compute/core/CL/kernels/CLYOLOLayerKernel.h
new file mode 100644
index 0000000..4c4aeac
--- /dev/null
+++ b/arm_compute/core/CL/kernels/CLYOLOLayerKernel.h
@@ -0,0 +1,86 @@
+/*
+ * 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_CLYOLOLAYERKERNEL_H__
+#define __ARM_COMPUTE_CLYOLOLAYERKERNEL_H__
+
+#include "arm_compute/core/CL/ICLKernel.h"
+
+namespace arm_compute
+{
+class ICLTensor;
+
+/** Interface for the YOLO layer kernel that performs partial activation.
+ * For each box, activate only:
+ * - x and y position (channel 0 and 1 of each box)
+ * - objectiveness (channel 4 of each box)
+ * - classes (channel 5 to (classes - 5) of each box)
+ */
+class CLYOLOLayerKernel : public ICLKernel
+{
+public:
+ /** Default constructor */
+ CLYOLOLayerKernel();
+ /** Prevent instances of this class from being copied (As this class contains pointers) */
+ CLYOLOLayerKernel(const CLYOLOLayerKernel &) = delete;
+ /** Prevent instances of this class from being copied (As this class contains pointers) */
+ CLYOLOLayerKernel &operator=(const CLYOLOLayerKernel &) = delete;
+ /** Allow instances of this class to be moved */
+ CLYOLOLayerKernel(CLYOLOLayerKernel &&) = default;
+ /** Allow instances of this class to be moved */
+ CLYOLOLayerKernel &operator=(CLYOLOLayerKernel &&) = default;
+ /** Default destructor */
+ ~CLYOLOLayerKernel() = default;
+ /** Set the input and output tensor.
+ *
+ * @note If the output tensor is a nullptr, the activation function will be performed in-place
+ *
+ * @param[in, out] input Source tensor. In case of @p output tensor = nullptr, this tensor will store the result
+ * of the activation function. Data types supported: F16/F32.
+ * @param[out] output Destination tensor. Data type supported: same as @p input
+ * @param[in] act_info Activation layer information.
+ * @param[in] num_classes Number of classes to activate (must be submultiple of @p input channels)
+ */
+ void configure(ICLTensor *input, ICLTensor *output, const ActivationLayerInfo &act_info, int32_t num_classes);
+ /** Static function to check if given info will lead to a valid configuration of @ref CLYOLOLayerKernel
+ *
+ * @param[in] input Source tensor info. In case of @p output tensor info = nullptr, this tensor will store the result
+ * of the activation function. Data types supported: F16/F32.
+ * @param[in] output Destination tensor info. Data type supported: same as @p input
+ * @param[in] act_info Activation layer information.
+ * @param[in] num_classes Number of classes to activate (must be submultiple of @p input channels)
+ *
+ * @return a status
+ */
+ static Status validate(const ITensorInfo *input, const ITensorInfo *output, const ActivationLayerInfo &act_info, int32_t num_classes);
+
+ // Inherited methods overridden:
+ void run(const Window &window, cl::CommandQueue &queue) override;
+
+private:
+ ICLTensor *_input;
+ ICLTensor *_output;
+ bool _run_in_place;
+};
+} // namespace arm_compute
+#endif /*__ARM_COMPUTE_CLYOLOLAYERKERNEL_H__ */
diff --git a/arm_compute/core/Utils.h b/arm_compute/core/Utils.h
index 1cdfd38..222f867 100644
--- a/arm_compute/core/Utils.h
+++ b/arm_compute/core/Utils.h
@@ -1011,6 +1011,12 @@
std::stringstream ss;
ss.precision(std::numeric_limits<float>::digits10 + 1);
ss << val;
+
+ if(val != static_cast<int>(val))
+ {
+ ss << "f";
+ }
+
return ss.str();
}