patch for rebase master on release/1.7.0submit/tizen/20200731.060745submit/tizen/20200730.075407submit/tizen/20200730.023729accepted/tizen/unified/20200803.122223tizen_6_m1_backup

Change-Id: Id38b617d325ef7e854995a47f032bdf482a779b3

15 files changed, 280 insertions, 1184 deletions

diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLArgOperation.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLArgOperation.h
deleted file mode 100644
index c37096f7c..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLArgOperation.h
+++ /dev/null
@@ -1,129 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * 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.
- */
-
-/**
- * @file CLArgOperation.h
- * @ingroup COM_AI_RUNTIME
- * @brief This file contains arm_compute::CLArgOperation class
- */
-
-#ifndef __ARM_COMPUTE_CLARGOPERATION_H__
-#define __ARM_COMPUTE_CLARGOPERATION_H__
-
-#include "arm_compute/core/CL/kernels/CLArgOperationKernel.h"
-#include "arm_compute/runtime/CL/CLTensor.h"
-#include "arm_compute/runtime/IFunction.h"
-#include "arm_compute/core/TypesEx.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/**
- * @brief Class to execute CLArgOperation operation
- */
-class CLArgOperation : public IFunction
-{
-public:
-  /**
-   * @brief Construct a new CLArgOperation object
-   */
-  CLArgOperation();
-
-  /**
-   * @brief Prevent instances of this class from being copied (As this class contains pointers)
-   */
-  CLArgOperation(const CLArgOperation &) = delete;
-
-  /**
-   * @brief Prevent instances of this class from being copied (As this class contains pointers)
-   */
-  CLArgOperation &operator=(const CLArgOperation &) = delete;
-
-  /**
-   * @brief Construct a new CLArgOperation object by using copy constructor
-   * @param[in] CLArgOperation object to move
-   */
-  CLArgOperation(CLArgOperation &&) = default;
-
-  /**
-   * @brief Assign a CLArgOperation object.
-   * @param[in] CLArgOperation object to assign. This object will be moved.
-   */
-  CLArgOperation &operator=(CLArgOperation &&) = default;
-
-  /**
-   * @brief Initialise the kernel's inputs and outputs.
-   * @param[in]  input     Input tensor. Data types supported: U8/QASYMM8/S32/F32.
-   * @param[out] output    The result of arg operation. Data types supported: S32.
-   * @param[in]  axis      Axis along which to reduce. It must be sorted and no duplicates.
-   * @param[in]  op        Arg operation to perform.
-   * @return N/A
-   */
-  void configure(ICLTensor *input, ICLTensor *output, std::vector<uint32_t> axis, ArgOperation op);
-
-  /**
-   * @brief Static function to check if given info will lead to a valid configuration
-   * @param[in]  input     Input tensor. Data types supported: U8/QASYMM8/S32/F32.
-   * @param[in]  axis      Axis along which to reduce. It must be sorted and no duplicates.
-   * @param[out] output    The result of arg operation. Data types supported: S32.
-   * @param[in]  op        Arg operation to perform.
-   * @return a status
-   */
-  static Status validate(const ITensorInfo *input, const std::vector<uint32_t> &axis,
-                         const ITensorInfo *output, ArgOperation op);
-  /**
-   * @brief Run the OpenCL kernel for this operation
-   * @return N/A
-   */
-  void run() override;
-
-private:
-  ICLTensor *_input{nullptr};
-  ICLTensor *_output{nullptr};
-  std::vector<uint32_t> _axis{};
-  ArgOperation _arg_op{ArgOperation::MAX};
-
-  std::unique_ptr<CLTensor[]> _interm_tensors{nullptr};
-  std::unique_ptr<CLArgOperationKernel[]> _argop_kernels{nullptr};
-  size_t _num_of_kernels{0};
-};
-}
-#endif /*__ARM_COMPUTE_CLARGOPERATION_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLBatchToSpaceND.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLBatchToSpaceND.h
deleted file mode 100644
index eed5cb8a4..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLBatchToSpaceND.h
+++ /dev/null
@@ -1,69 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2016-2018 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_CLBATCH_TO_SPACE_ND_H__
-#define __ARM_COMPUTE_CLBATCH_TO_SPACE_ND_H__
-
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Basic function to run @ref CLBatchToSpaceNDKernel
- *
- * @note The tensor data type for the inputs must be U8/QASYMM8/S16/S32/F16/F32.
- * @note The function converts the input tensor to the tensor of the output tensor's type.
- */
-class CLBatchToSpaceND : public ICLSimpleFunction
-{
-public:
-  /** Initialise the kernel's input and output.
-   *
-   * @param[in]  input              Input tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.
-   * @param[out] output             Output tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.
-   * @param[in]  block_size         A pointer to an array of integer values specifying block sizes
-   *                                for spatial dimension.
-   */
-  void configure(ICLTensor *input, ICLTensor *output, const int32_t *block_size);
-};
-
-} // namespace arm_compute
-#endif /* __ARM_COMPUTE_CLBATCH_TO_SPACE_ND_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLCast.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLCast.h
deleted file mode 100644
index ebe0d8a1c..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLCast.h
+++ /dev/null
@@ -1,75 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2016-2018 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.
- */
-
-/**
- * @file CLCast.h
- * @ingroup COM_AI_RUNTIME
- * @brief This file contains arm_compute::CLCast class
- */
-
-#ifndef __ARM_COMPUTE_CLCAST_H__
-#define __ARM_COMPUTE_CLCAST_H__
-
-#include "arm_compute/core/TypesEx.h"
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/**
- * @brief Class to run @ref CLCastKernel.
- * This converts the input tensor to the tensor of the output tensor's type.
- */
-class CLCast : public ICLSimpleFunction
-{
-public:
-  /**
-   * @brief Initialise the kernel's input and output
-   * @param[in, out] input    Input tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.
-   *                          The input tensor is [in, out] because its TensorInfo might be
-   *                          modified inside the kernel.
-   * @param[out]     output   Output tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.
-   * @param[in]      input_subtype  Sub data type of input.
-   */
-  void configure(ICLTensor *input, ICLTensor *output, SubDataType input_subtype);
-};
-}
-#endif /* __ARM_COMPUTE_CLCAST_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLDepthToSpace.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLDepthToSpace.h
deleted file mode 100644
index d52a538df..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLDepthToSpace.h
+++ /dev/null
@@ -1,68 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2016-2018 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_CLDEPTHTOSPACE_H__
-#define __ARM_COMPUTE_CLDEPTHTOSPACE_H__
-
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Basic function to run @ref CLDepthToSpaceKernel
- *
- * @note The tensor data type for the inputs must be U8/QASYMM8/S16/S32/F16/F32.
- * @note The function converts the input tensor to the tensor of the output tensor's type.
- */
-class CLDepthToSpace : public ICLSimpleFunction
-{
-public:
-  /** Initialise the kernel's input and output.
-   *
-   * @param[in]  input              Input tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.
-   * @param[out] output             Output tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.
-   * @param[block_size] block size  integer only
-   */
-  void configure(ICLTensor *input, ICLTensor *output, const int32_t block_size);
-};
-} // namesace arm_compute
-
-#endif /* __ARM_COMPUTE_CLDEPTHTOSPACE_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLDirectTransposeConvLayer.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLDirectTransposeConvLayer.h
new file mode 100644
index 000000000..409eaf593
--- /dev/null
+++ b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLDirectTransposeConvLayer.h
@@ -0,0 +1,201 @@
+/*
+ * Copyright (c) 2020 Samsung Electronics Co., Ltd. All Rights Reserved
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *      http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/*
+ * Copyright (c) 2019-2020 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_CLDIRECTTRANSPOSECONVLAYER_H__
+#define __ARM_COMPUTE_CLDIRECTTRANSPOSECONVLAYER_H__
+
+#include "arm_compute/runtime/CL/functions/CLConvolutionLayer.h"
+#include "arm_compute/runtime/CL/functions/CLDeconvolutionLayerUpsample.h"
+#include "arm_compute/runtime/CL/functions/CLReverse.h"
+#include "arm_compute/runtime/CL/functions/CLTranspose.h"
+
+#include "arm_compute/runtime/CL/CLTensor.h"
+#include "arm_compute/runtime/IFunction.h"
+#include "arm_compute/runtime/IMemoryManager.h"
+#include "arm_compute/runtime/MemoryGroup.h"
+
+#include <memory>
+
+namespace arm_compute
+{
+class ICLTensor;
+/** Function to run the deconvolution layer.
+ *
+ * Deconvolution Layer is the backward pass of Convolution Layer. First we transform the input
+ * depending on the stride and pad info and then perform a 1x1
+ * convolution pass. Input stride defines how many zeroes we should put between each element of the
+ * input and pad is the amount of padding.
+ *
+ *  The relation between input to output is as follows:
+ *  \f[
+ *       width\_output = (width\_input - 1) \cdot stride\_x - 2 \cdot padding\_x + kernel\_x
+ *  \f]
+ *  \f[
+ *       height\_output = (height\_input - 1) \cdot stride\_y - 2 \cdot padding\_y + kernel\_y
+ *  \f]
+ *
+ *  where:
+ *      width_input is the size of the first input dimension.
+ *      height_input 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.
+ *      stride_x and stride_y is the input stride of the first and second dimension.
+ *
+ * The weights used by Deconvolution are supposed to be the same as the ones used for Convolution.
+ * Therefore, it will be necessary to use the weights in the
+ * reverse order to perform an actual convolution. This is achieved by using @ref CLReverse.
+ *
+ * This function calls the following OpenCL kernels/functions:
+ *
+ * -# @ref CLDeconvolutionLayerUpsample
+ * -# @ref CLConvolutionLayer
+ *
+ * And the following CPP kernels:
+ * -# @ref CLReverse
+ *
+ */
+class CLDirectTransposeConvLayer : public IFunction
+{
+public:
+  /** Constructor */
+  CLDirectTransposeConvLayer(std::shared_ptr<IMemoryManager> memory_manager = nullptr);
+  /** Prevent instances of this class from being copied (As this class contains pointers) */
+  CLDirectTransposeConvLayer(const CLDirectTransposeConvLayer &) = delete;
+  /** Default move constructor */
+  CLDirectTransposeConvLayer(CLDirectTransposeConvLayer &&) = default;
+  /** Prevent instances of this class from being copied (As this class contains pointers) */
+  CLDirectTransposeConvLayer &operator=(const CLDirectTransposeConvLayer &) = delete;
+  /** Default move assignment operator */
+  CLDirectTransposeConvLayer &operator=(CLDirectTransposeConvLayer &&) = default;
+  /** 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: QASYMM8_SIGNED/QASYMM8/F16/F32.
+   * @param[in]     weights      The 4d weights with dimensions [width, height, IFM, OFM]. Data type
+ * supported: Same as @p input.
+   * @param[in]     bias         (Optional) The biases have one dimension.
+   *                             Data type supported: Should match @p input data type, except for
+ * input of QASYMM8 and QASYMM8_SIGNED type where biases should be of S32 type
+   * @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] invalid_right  The number of zeros added to right edge of the output.
+ * @param[in] invalid_bottom  The number of zeros added to bottom edge of the output.
+   * @param[in]     weights_info (Optional) Weights information needed for @ref CLConvolutionLayer,
+ * specifies if the weights tensor has been reshaped with @ref CLWeightsReshapeKernel.
+   *
+   */
+  void configure(ICLTensor *input, ICLTensor *weights, const ICLTensor *bias, ICLTensor *output,
+                 const PadStrideInfo &info, unsigned int invalid_right, unsigned int invalid_bottom,
+                 const WeightsInfo &weights_info = WeightsInfo());
+  /** Set the input, weights, biases and output tensors.
+   *
+   * @param[in]     compile_context The compile context to be used.
+   * @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: QASYMM8_SIGNED/QASYMM8/F16/F32.
+   * @param[in]     weights         The 4d weights with dimensions [width, height, IFM, OFM]. Data
+ * type supported: Same as @p input.
+   * @param[in]     bias            (Optional) The biases have one dimension.
+   *                                Data type supported: Should match @p input data type, except for
+ * input of QASYMM8 and QASYMM8_SIGNED type where biases should be of S32 type
+   * @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] invalid_right  The number of zeros added to right edge of the output.
+ * @param[in] invalid_bottom  The number of zeros added to bottom edge of the output.
+   * @param[in]     weights_info    (Optional) Weights information needed for @ref
+ * CLConvolutionLayer, specifies if the weights tensor has been reshaped with @ref
+ * CLWeightsReshapeKernel.
+   *
+   */
+  void configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *weights,
+                 const ICLTensor *bias, ICLTensor *output, const PadStrideInfo &info,
+                 unsigned int invalid_right, unsigned int invalid_bottom,
+                 const WeightsInfo &weights_info = WeightsInfo());
+  /** Static function to check if given info will lead to a valid configuration of @ref
+ * CLDirectTransposeConvLayer
+   *
+   * @param[in] input        Input tensor info. 3 lower dimensions represent a single input, and an
+ * optional 4th dimension for batch of inputs.
+   *                         Data types supported: QASYMM8_SIGNED/QASYMM8/F16/F32.
+   * @param[in] weights      The 4d weights info with dimensions [width, height, IFM, OFM]. Data
+ * type supported: Same as @p input.
+   * @param[in] bias         (Optional) The biases have one dimension.
+   *                         Data type supported: Should match @p input data type, except for input
+ * of QASYMM8 and QASYMM8_SIGNED type where biases should be of S32 type
+   * @param[in] output       Output tensor info. 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] invalid_right  The number of zeros added to right edge of the output.
+ * @param[in] invalid_bottom  The number of zeros added to bottom edge of the output.
+   * @param[in] weights_info (Optional) Weights information needed for @ref CLConvolutionLayer,
+ * specifies if the weights tensor has been reshaped with @ref CLWeightsReshapeKernel.
+   *
+   * @return a status
+   */
+  static Status validate(const ITensorInfo *input, const ITensorInfo *weights,
+                         const ITensorInfo *bias, ITensorInfo *output, const PadStrideInfo &info,
+                         unsigned int invalid_right, unsigned int invalid_bottom,
+                         const WeightsInfo &weights_info = WeightsInfo());
+
+  // Inherited methods overridden:
+  void run() override;
+  void prepare() override;
+
+private:
+  MemoryGroup _memory_group;
+  CLDeconvolutionLayerUpsample _scale_f;
+  CLConvolutionLayer _conv_f;
+  CLReverse _flip_weights;
+
+  CLTensor _scaled_output;
+  ICLTensor *_original_weights;
+  CLTensor _weights_flipped;
+  CLTensor _flip_axis;
+
+  bool _is_prepared;
+};
+} // namespace arm_compute
+#endif /* __ARM_COMPUTE_CLDIRECTTRANSPOSECONVLAYER_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLFullyConnectedHybridLayer.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLFullyConnectedHybridLayer.h
index 1a0284a3e..f3266f688 100644
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLFullyConnectedHybridLayer.h
+++ b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLFullyConnectedHybridLayer.h
@@ -50,7 +50,7 @@
 #include "arm_compute/core/CL/kernels/CLTransposeKernel.h"
 #include "arm_compute/runtime/MemoryGroup.h"
 #include "arm_compute/runtime/CL/CLTensor.h"
-#include "arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCoreEx.h"
+#include "arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCore.h"
 
 namespace arm_compute
 {
@@ -168,7 +168,7 @@ private:
   CLFullyConnectedHybridLayerReshapeWeights _reshape_weights_kernel;
   CLScaleFactorSymm8Kernel _scale_factor_kernel;
   CLQuantizationSymmetricKernel _quant_input_kernel;
-  CLGEMMLowpMatrixMultiplyCoreEx _mm_gemmlowp;
+  CLGEMMLowpMatrixMultiplyCore _mm_gemmlowp;
   CLMultiplyScaleFactorKernel _multiply_scale_kernel;
   CLGEMMMatrixAccumulateBiasesKernel _accumulate_biases_kernel; // TODO(COMPMID-1889): Use CLGEMM to
                                                                 // add bias in
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCoreEx.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCoreEx.h
deleted file mode 100644
index 68aba74ab..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLGEMMLowpMatrixMultiplyCoreEx.h
+++ /dev/null
@@ -1,142 +0,0 @@
-/*
- * Copyright (c) 2020 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2017-2019 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_CLGEMMLOWPMATRIXMULTIPLYCOREEX_H__
-#define __ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYCOREEX_H__
-
-#include "arm_compute/core/CL/kernels/CLDepthConvertLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyKernelEx.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMLowpReductionKernel.h"
-#include "arm_compute/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.h"
-#include "arm_compute/runtime/CL/CLTensor.h"
-#include "arm_compute/runtime/IFunction.h"
-#include "arm_compute/runtime/MemoryGroup.h"
-
-namespace arm_compute
-{
-class IMemoryManager;
-class ICLTensor;
-
-/** Basic function to execute GEMMLowpMatrixMultiplyCore on OpenCL. This function calls the
- * following OpenCL kernels:
- *
- *  -# @ref CLGEMMLowpMatrixMultiplyKernel (if the parameter "reshape_b_only_on_first_run" of
- * GEMMInfo is FALSE)
- *  -# @ref CLGEMMLowpMatrixAReductionKernel (if the offset of matrix B is not 0)
- *  -# @ref CLGEMMLowpMatrixBReductionKernel (if the offset of matrix A is not 0)
- *
-*/
-class CLGEMMLowpMatrixMultiplyCoreEx : public IFunction
-{
-public:
-  /** Constructor */
-  CLGEMMLowpMatrixMultiplyCoreEx(std::shared_ptr<IMemoryManager> memory_manager = nullptr);
-  /** Prevent instances of this class from being copied (As this class contains pointers) */
-  CLGEMMLowpMatrixMultiplyCoreEx(const CLGEMMLowpMatrixMultiplyCoreEx &) = delete;
-  /** Default move constructor */
-  CLGEMMLowpMatrixMultiplyCoreEx(CLGEMMLowpMatrixMultiplyCoreEx &&) = default;
-  /** Prevent instances of this class from being copied (As this class contains pointers) */
-  CLGEMMLowpMatrixMultiplyCoreEx &operator=(const CLGEMMLowpMatrixMultiplyCoreEx &) = delete;
-  /** Default move assignment operator */
-  CLGEMMLowpMatrixMultiplyCoreEx &operator=(CLGEMMLowpMatrixMultiplyCoreEx &&) = default;
-  /** Initialise the kernel's inputs, output
-   *
-   * @note GEMMLowp:  low precision GEMM kernel. [A * B + C]
-   *  This kernel performs the following computations:
-   *
-   *  -# Convert a values from QASYMM8 to int32 and add a_offset to each of them.
-   *  -# Convert b values from QASYMM8 to int32 and add b_offset to each of them.
-   *  -# Compute the matrix product of the resulting a * b in int32.
-   *  -# Quantize to uint8 if gemm_info.gemmlowp_output_stage != NONE
-   *
-   * @param[in]  a         First input tensor  (Matrix A). Data type supported: QASYMM8.
-   * @param[in]  b         Second input tensor (Matrix B). Data type supported: same as @p a
-   * @param[in]  c         Third input tensor  (Matrix C). It can be a nullptr. Data type supported:
-   * S32
-   * @param[out] output    Output tensor. Data type supported: S32 or QASYMM8 if
-   * gemm_info.gemmlowp_output_stage != NONE
-   * @param[in]  gemm_info (Optional) Specifies if the matrix A and/or matrix B have been reshaped
-   * and
-   *                       if the reshape of matrix B should be executed only for the first run
-   */
-  void configure(const ICLTensor *a, const ICLTensor *b, const ICLTensor *c, ICLTensor *output,
-                 const GEMMInfo &gemm_info = GEMMInfo());
-  /** Static function to check if given info will lead to a valid configuration of @ref
-   * CLGEMMLowpMatrixMultiplyCoreEx
-   *
-   * @param[in] a         First input tensor info (Matrix A). Data type supported: QASYMM8.
-   * @param[in] b         Second input tensor info (Matrix B). Data type supported: same as @p a
-   * @param[in] c         Third input tensor info (Matrix C). It can be a nullptr. Data type
-   * supported: S32
-   * @param[in] output    Output tensor info. Data type supported: S32 or QASYMM8 if
-   * gemm_info.gemmlowp_output_stage != NONE
-   * @param[in] gemm_info (Optional) Specifies if the matrix A and/or matrix B have been reshaped
-   * and
-   *                      if the reshape of matrix B should be executed only for the first run
-   *
-   * @return a status
-   */
-  static Status validate(const ITensorInfo *a, const ITensorInfo *b, const ITensorInfo *c,
-                         const ITensorInfo *output, const GEMMInfo &gemm_info = GEMMInfo());
-
-  // Inherited methods overridden:
-  void run() override;
-  void prepare() override;
-
-private:
-  MemoryGroup _memory_group;
-
-  // Kernels used
-  CLGEMMLowpMatrixMultiplyKernelEx _mm_midgard_kernel;
-  CLGEMMLowpMatrixAReductionKernel _mtx_a_reduction_kernel;
-  CLGEMMLowpMatrixBReductionKernel _mtx_b_reduction_kernel;
-
-  // Temporary tensors
-  CLTensor _vector_sum_col;
-  CLTensor _vector_sum_row;
-
-  int32_t _a_offset;
-  int32_t _b_offset;
-  bool _reshape_b_only_on_first_run;
-  bool _is_prepared;
-};
-} // namespace arm_compute
-#endif /*__ARM_COMPUTE_CLGEMMLOWPMATRIXMULTIPLYCOREEX_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLLogicalNot.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLLogicalNot.h
deleted file mode 100644
index 51216715f..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLLogicalNot.h
+++ /dev/null
@@ -1,62 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2016-2018 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_CLLOGICALNOT_H__
-#define __ARM_COMPUTE_CLLOGICALNOT_H__
-
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-class CLLogicalNot : public ICLSimpleFunction
-{
-public:
-  /** Initialise the function's source and destination.
-   *
-   * @param[in]  input  Source tensor. Data types supported: QASYMM8.
-   * @param[out] output Output tensor. Data types supported: QASYMM8.
-   */
-  void configure(ICLTensor *input, ICLTensor *output);
-};
-
-} // namespace arm_compute
-#endif /*__ARM_COMPUTE_CLLOGICALNOT_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLPReLU.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLPReLU.h
deleted file mode 100644
index 7fbe558ff..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLPReLU.h
+++ /dev/null
@@ -1,64 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2016-2018 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_CLPRELU_H__
-#define __ARM_COMPUTE_CLPRELU_H__
-
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-class CLPReLU : public ICLSimpleFunction
-{
-public:
-  /** Initialise the function's source and destination.
-   *
-   * @param[in]  input. Data types supported:
-   * QASYMM8/F16/F32.
-   * @param[in]  alpha. Data types supported:
-   * QASYMM8/F16/F32.
-   * @param[out] output Output tensor. Data types supported: Same as @p input.
-   */
-  void configure(ICLTensor *input, ICLTensor *alpha, ICLTensor *output);
-};
-} // namespace arm_compute
-#endif /*__ARM_COMPUTE_CLPRELU_H__*/
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLPixelWiseDivision.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLPixelWiseDivision.h
deleted file mode 100644
index e83fb01cd..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLPixelWiseDivision.h
+++ /dev/null
@@ -1,103 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2016-2018 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.
- */
-
-/**
- * @file CLPixelWiseDivision.h
- * @ingroup COM_AI_RUNTIME
- * @brief This file contains arm_compute::CLPixelWiseDivision class
- */
-#ifndef __ARM_COMPUTE_CLPIXELWISEDIVISION_H__
-#define __ARM_COMPUTE_CLPIXELWISEDIVISION_H__
-
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/**
- * @brief Class to run @ref CLPixelWiseDivisionKernel.
- */
-class CLPixelWiseDivision : public ICLSimpleFunction
-{
-public:
-  /**
-   * @brief Initialise the kernel's inputs, output and convertion policy.
-   * @param[in, out] input1          An input tensor. Data types supported: U8/S16/F16/F32
-   *                                 The input tensor is [in, out] because its TensorInfo might be
-   * modified inside the kernel in case of broadcasting of dimension 0.
-   * @param[in, out] input2          An input tensor. Data types supported: same as @p input1.
-   *                                 The input tensor is [in, out] because its TensorInfo might be
-   * modified inside the kernel in case of broadcasting of dimension 0.
-   * @param[out]     output          The output tensor, Data types supported: same as @p input1.
-   * Note: U8 requires both inputs to be U8.
-   * @param[in]      scale           Scale to apply after multiplication.
-   *                                 Scale must be positive and its value must be either 1/255 or
-   * 1/2^n where n is between 0 and 15.
-   * @param[in]      overflow_policy Overflow policy. Supported overflow policies: Wrap, Saturate
-   * @param[in]      rounding_policy Rounding policy. Supported rounding modes: to zero, to nearest
-   * even.
-   * @return N/A
-   */
-  void configure(ICLTensor *input1, ICLTensor *input2, ICLTensor *output, float scale = 1.f,
-                 ConvertPolicy overflow_policy = ConvertPolicy::WRAP,
-                 RoundingPolicy rounding_policy = RoundingPolicy::TO_ZERO);
-
-  /**
-   * @brief Static function to check if given info will lead to a valid configuration of @ref
-   * CLPixelWiseDivision
-   * @param[in] input1          An input tensor info. Data types supported: U8/S16/F16/F32
-   * @param[in] input2          An input tensor info. Data types supported: same as @p input1.
-   * @param[in] output          The output tensor info, Data types supported: same as @p input1.
-   * Note: U8 requires both inputs to be U8.
-   * @param[in] scale           Scale to apply after multiplication.
-   *                            Scale must be positive and its value must be either 1/255 or 1/2^n
-   * where n is between 0 and 15.
-   * @param[in] overflow_policy Overflow policy. Supported overflow policies: Wrap, Saturate
-   * @param[in] rounding_policy Rounding policy. Supported rounding modes: to zero, to nearest even.
-   * @return a status
-   */
-  static Status validate(const ITensorInfo *input1, const ITensorInfo *input2,
-                         const ITensorInfo *output, float scale = 1.f,
-                         ConvertPolicy overflow_policy = ConvertPolicy::WRAP,
-                         RoundingPolicy rounding_policy = RoundingPolicy::TO_ZERO);
-};
-}
-#endif /*__ARM_COMPUTE_CLPIXELWISEDIVISION_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLRNNLayerEx.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLRNNLayerEx.h
deleted file mode 100644
index b49cbd873..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLRNNLayerEx.h
+++ /dev/null
@@ -1,120 +0,0 @@
-/*
- * Copyright (c) 2019 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * 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_CLRNN_LAYER_EX_H__
-#define __ARM_COMPUTE_CLRNN_LAYER_EX_H__
-
-#include "arm_compute/core/CL/kernels/CLActivationLayerKernel.h"
-#include "arm_compute/core/CL/kernels/CLCopyKernel.h"
-#include "arm_compute/core/CL/kernels/CLElementwiseOperationKernel.h"
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-#include "arm_compute/runtime/CL/functions/CLFullyConnectedLayer.h"
-#include "arm_compute/runtime/CL/functions/CLGEMM.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Basic function to run @ref CLRNNLayerEx */
-class CLRNNLayerEx : public IFunction
-{
-public:
-  /** Default constructor */
-  CLRNNLayerEx(std::shared_ptr<IMemoryManager> memory_manager = nullptr);
-  /** Initialize the function
-   *
-   * @param[in]     input             Input is a 2-D tensor of shape [input_size, batch_size]. Data
-   * types supported: F16/F32
-   * @param[in]     weights           Weights tensor of shape [input_size, num_units] that
-   * multiplies the input. Data types supported: Same as @p input
-   * @param[in]     recurrent_weights Weights tensor of shape [num_units, num_units] that multiplies
-   * the current 'state'. Data types supported: Same as @p input
-   * @param[in]     bias              Bias vector of shape [num_units]. Data types supported: Same
-   * as @p input
-   * @param[out]    output            Output tensor of shape [num_units, batch_size]. Data types
-   * supported: Same as @p input
-   * @param[in,out] hidden_state      Output tensor of shape [num_units, batch_size]. Data types
-   * supported: Same as @p input
-   * @param[in]     info              Activation layer parameter.
-   */
-  void configure(const ICLTensor *input, const ICLTensor *weights,
-                 const ICLTensor *recurrent_weights, const ICLTensor *bias, ICLTensor *hidden_state,
-                 ICLTensor *output, ActivationLayerInfo &info);
-  /** Initialize the function
-   *
-   * @param[in] input             Input is a 2-D tensor of shape [input_size, batch_size]. Data
-   * types supported: F16/F32
-   * @param[in] weights           Weights tensor of shape [input_size, num_units] that multiplies
-   * the input. Data types supported: Same as @p input
-   * @param[in] recurrent_weights Weights tensor of shape [num_units, num_units] that multiplies the
-   * current 'state'. Data types supported: Same as @p input
-   * @param[in] bias              Bias vector of shape [num_units]. Data types supported: Same as @p
-   * input
-   * @param[in] output            Output tensor of shape [num_units, batch_size]. Data types
-   * supported: Same as @p input
-   * @param[in] hidden_state      Output tensor of shape [num_units, batch_size]. Data types
-   * supported: Same as @p input
-   * @param[in] info              Activation layer parameter.
-   *
-   * @return a status
-   */
-  static Status validate(const ITensorInfo *input, const ITensorInfo *weights,
-                         const ITensorInfo *recurrent_weights, const ITensorInfo *bias,
-                         const ITensorInfo *hidden_state, const ITensorInfo *output,
-                         const ActivationLayerInfo &info);
-
-  // Inherited methods overridden:
-  void run() override;
-  void prepare() override;
-
-private:
-  MemoryGroup _memory_group;
-  CLGEMM _gemm_state_f;
-  CLSaturatedArithmeticOperationKernel _add_kernel;
-  CLActivationLayerKernel _activation_kernel;
-  CLFullyConnectedLayer _fully_connected_kernel;
-  CLCopyKernel _copy_kernel;
-  CLTensor _fully_connected_out;
-  CLTensor _gemm_output;
-  CLTensor _add_output;
-  bool _is_prepared;
-};
-}
-#endif /* __ARM_COMPUTE_CLRNN_LAYER_EX_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLSpaceToDepth.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLSpaceToDepth.h
deleted file mode 100644
index 2090b46fa..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLSpaceToDepth.h
+++ /dev/null
@@ -1,68 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2016-2018 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_CLSPACETODEPTH_H__
-#define __ARM_COMPUTE_CLSPACETODEPTH_H__
-
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Basic function to run @ref CLSpaceToDepthKernel
- *
- * @note The tensor data type for the inputs must be U8/QASYMM8/S16/S32/F16/F32.
- * @note The function converts the input tensor to the tensor of the output tensor's type.
- */
-class CLSpaceToDepth : public ICLSimpleFunction
-{
-public:
-  /** Initialise the kernel's input and output.
-   *
-   * @param[in]  input              Input tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.
-   * @param[out] output             Output tensor. Data types supported: U8/QASYMM8/S16/S32/F16/F32.
-   * @param[block_size] block size  integer only
-   */
-  void configure(ICLTensor *input, ICLTensor *output, const int32_t block_size);
-};
-
-} // namespace arm_compute
-#endif /* __ARM_COMPUTE_CLSPACETODEPTH_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLStridedSliceEx.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLStridedSliceEx.h
deleted file mode 100644
index 03edd15e6..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLStridedSliceEx.h
+++ /dev/null
@@ -1,81 +0,0 @@
-/*
- * Copyright (c) 2018 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * 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.
- */
-
-/**
- * @file CLStridedSlice.h
- * @ingroup COM_AI_RUNTIME
- * @brief This file contains arm_compute::CLStridedSlice and arm_compute::CLStridedSliceCPU class
- */
-
-#ifndef __ARM_COMPUTE_CLSTRIDEDSLICEEX_H__
-#define __ARM_COMPUTE_CLSTRIDEDSLICEEX_H__
-
-#include "arm_compute/runtime/CL/ICLSimpleFunction.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/**
- * @brief Class to run @ref CLStridedSliceKernel
- */
-class CLStridedSliceEx : public ICLSimpleFunction
-{
-public:
-  /**
-   * @brief Initialise the kernel's inputs and outputs
-   * @param[in]  input   Tensor input. Data type supported:
-   *                     U8/S8/QASYMM8/U16/S16/U32/S32/F16/F32
-   * @param[out] output  Output tensor. Data type supported: Same as @p input
-   * @param[in]  beginData 'begin' vector of strided slice operation
-   * @param[in]  endData   'end' vector of strided slice operation
-   * @param[in]  stridesData 'strides' vector of strided slice operation
-   * @param[in]  beginMask  If the ith bit is set, begin[i] is ignored
-   * @param[in]  endMask    If the ith bit is set, end[i] is ignored
-   * @param[in]  shrinkAxisMask  If the ith bit is set, the ith specification shrinks the
-   *                             dimensionality by 1, taking on the value at index begin[i]
-   * @return N/A
-   */
-  void configure(const ICLTensor *input, ICLTensor *output, ICLTensor *beginData,
-                 ICLTensor *endData, ICLTensor *stridesData, int32_t beginMask, int32_t endMask,
-                 int32_t shrinkAxisMask);
-};
-}
-#endif /*__ARM_COMPUTE_CLSTRIDEDSLICEEX_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLTransposeConvLayer.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLTransposeConvLayer.h
index 54a697e69..5fb102e47 100644
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLTransposeConvLayer.h
+++ b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLTransposeConvLayer.h
@@ -15,7 +15,7 @@
  */
 
 /*
- * Copyright (c) 2017-2018 ARM Limited.
+ * Copyright (c) 2017-2020 ARM Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -37,16 +37,11 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
  * SOFTWARE.
  */
-
 #ifndef __ARM_COMPUTE_CLTRANSPOSECONVLAYER_H__
 #define __ARM_COMPUTE_CLTRANSPOSECONVLAYER_H__
 
-#include "arm_compute/runtime/CL/functions/CLConvolutionLayer.h"
-#include "arm_compute/runtime/CL/functions/CLTransposeConvLayerUpsample.h"
-
-#include "arm_compute/core/CPP/kernels/CPPFlipWeightsKernel.h"
-
-#include "arm_compute/runtime/CL/CLTensor.h"
+#include "arm_compute/runtime/CL/functions/CLDirectTransposeConvLayer.h"
+#include "arm_compute/runtime/CL/functions/CLGEMMDeconvolutionLayer.h"
 #include "arm_compute/runtime/IFunction.h"
 #include "arm_compute/runtime/IMemoryManager.h"
 
@@ -54,119 +49,102 @@
 
 namespace arm_compute
 {
-class ICLTensor;
-/** Function to run the transpose convolution layer.
- *
- * @note This layer was copied in order to fix a bug computing to wrong output dimensions.
- *
- * TransposeConv Layer is the backward pass of Convolution Layer. First we transform the input
- * depending on the stride and pad info and then perform a 1x1
- * convolution pass. Input stride defines how many zeroes we should put between each element of the
- * input, pad is the amount of padding and finally a is a user
- * specified value where a < stride - 1, that increases the padding top and right of the input
- * image.
- *
- *  The relation between input to output is as follows:
- *  \f[
- *       width\_output = (width\_input - 1) \cdot stride\_x - \cdot padding\_x + kernel\_x
- *  \f]
- *  \f[
- *       height\_output = (height\_input - 1) \cdot stride\_y - \cdot padding\_y + kernel\_y
- *  \f]
- *
- *  where:
- *      width_input is the size of the first input dimension.
- *      height_input 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.
- *      stride_x and stride_y is the input stride of the first and second dimension.
- *
- * The weights used by Deconvolution are supposed to be the same as the ones used for Convolution.
- * Therefore, it will be necessary to use the weights in the
- * reverse order to perform an actual convolution. This is achieved by using the @ref
- * CPPFlipWeightsKernel.
- *
- * This function calls the following OpenCL kernels/functions:
- *
- * -# @ref CLTransposeConvLayerUpsample
- * -# @ref CLConvolutionLayer
+/** Basic function to compute the deconvolution layer. This function calls the following OpenCL
+ * kernels/functions:
  *
+ * -# @ref CLGEMMDeconvolutionLayer
+ * -# @ref CLDirectTransposeConvLayer
  */
 class CLTransposeConvLayer : public IFunction
 {
 public:
-  /** Constructor */
+  /** Default constructor */
   CLTransposeConvLayer(std::shared_ptr<IMemoryManager> memory_manager = nullptr);
-  /** Prevent instances of this class from being copied (As this class contains pointers) */
-  CLTransposeConvLayer(const CLTransposeConvLayer &) = delete;
-  /** Default move constructor */
-  CLTransposeConvLayer(CLTransposeConvLayer &&) = default;
-  /** Prevent instances of this class from being copied (As this class contains pointers) */
-  CLTransposeConvLayer &operator=(const CLTransposeConvLayer &) = delete;
-  /** Default move assignment operator */
-  CLTransposeConvLayer &operator=(CLTransposeConvLayer &&) = default;
+
   /** 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: QASYMM8/F16/F32.
-   * @param[in]     weights        The 4d weights with dimensions [width, height, IFM, OFM].
-   *                               Data type supported: Same as @p input.
-   * @param[in]     bias           (Optional) 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
-   *                               transpose convolution, this is decribed in @ref PadStrideInfo.
-   * @param[in]     invalid_right  The number of zeros added to right edge of the output.
-   * @param[in]     invalid_bottom The number of zeros added to top edge of the output.
-   * @param[in]     weights_info   (Optional) Weights information needed for @ref
-   *                               CLConvolutionLayer, specifies if the weights tensor has been
-   *                               reshaped with @ref CLWeightsReshapeKernel.
+   * @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: QASYMM8_SIGNED/QASYMM8/F16/F32.
+   * @param[in]     weights      The 4d weights with dimensions [width, height, IFM, OFM]. Data type
+ * supported: Same as @p input.
+   * @param[in]     bias         (Optional) 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]     deconv_info  Contains padding and policies to be used in the deconvolution, this
+ * is described in @ref PadStrideInfo.
+ * @param[in] invalid_right  The number of zeros added to right edge of the output.
+ * @param[in] invalid_bottom  The number of zeros added to bottom edge of the output.
+   * @param[in]     weights_info (Optional) Weights information needed for @ref CLConvolutionLayer,
+ * specifies if the weights tensor has been reshaped with @ref CLWeightsReshapeKernel.
+   *
    */
   void configure(ICLTensor *input, ICLTensor *weights, const ICLTensor *bias, ICLTensor *output,
-                 const PadStrideInfo &info, unsigned int invalid_right, unsigned int invalid_bottom,
+                 const PadStrideInfo &deconv_info, unsigned int invalid_right,
+                 unsigned int invalid_bottom, const WeightsInfo &weights_info = WeightsInfo());
+  /** Set the input, weights, biases and output tensors.
+   *
+   * @param[in]     compile_context The compile context to be used.
+   * @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:
+ * QASYMM8_SIGNED/QASYMM8/F16/F32.
+   * @param[in]     weights         The 4d weights with dimensions [width, height, IFM, OFM]. Data
+ * type supported: Same as @p input.
+   * @param[in]     bias            (Optional) 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]     deconv_info     Contains padding and policies to be used in the deconvolution,
+ * this is described in @ref PadStrideInfo.
+ * @param[in] invalid_right  The number of zeros added to right edge of the output.
+ * @param[in] invalid_bottom  The number of zeros added to bottom edge of the output.
+   * @param[in]     weights_info    (Optional) Weights information needed for @ref
+ * CLConvolutionLayer, specifies if the weights tensor has been reshaped with @ref
+ * CLWeightsReshapeKernel.
+   *
+   */
+  void configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *weights,
+                 const ICLTensor *bias, ICLTensor *output, const PadStrideInfo &deconv_info,
+                 unsigned int invalid_right, unsigned int invalid_bottom,
                  const WeightsInfo &weights_info = WeightsInfo());
   /** Static function to check if given info will lead to a valid configuration of @ref
-   * CLTransposeConvLayer
+ * CLTransposeConvLayer
+   *
+   * @param[in] input        Input tensor info. 3 lower dimensions represent a single input, and an
+ * optional 4th dimension for batch of inputs. Data types supported: QASYMM8_SIGNED/QASYMM8/F16/F32.
+   * @param[in] weights      The 4d weights info with dimensions [width, height, IFM, OFM]. Data
+ * type supported: Same as @p input.
+   * @param[in] bias         (Optional) The biases have one dimension. Data type supported: Same as
+ * @p input.
+   * @param[in] output       Output tensor info. The output has the same number of dimensions as the
+ * @p input.
+   * @param[in] deconv_info  Contains padding and policies to be used in the deconvolution, this is
+ * described in @ref PadStrideInfo.
+ * @param[in] invalid_right  The number of zeros added to right edge of the output.
+ * @param[in] invalid_bottom  The number of zeros added to bottom edge of the output.
+   * @param[in] weights_info (Optional) Weights information needed for @ref CLConvolutionLayer,
+ * specifies if the weights tensor has been reshaped with @ref CLWeightsReshapeKernel.
    *
-   * @param[in] input           Input tensor info. 3 lower dimensions represent a single input,
-   *                            and an optional 4th dimension for batch of inputs.
-   *                            Data types supported: QASYMM8/F16/F32.
-   * @param[in] weights         The 4d weights info with dimensions [width, height, IFM, OFM].
-   *                            Data type supported: Same as @p input.
-   * @param[in] bias            (Optional) The biases have one dimension. Data type supported:
-   *                            Same as @p input.
-   * @param[in] output          Output tensor info. The output has the same number of dimensions
-   *                            as the @p input.
-   * @param[in] info            Contains padding and policies to be used in the
-   *                            transpose convolution, this is decribed in @ref PadStrideInfo.
-   * @param[in] innvalid_right  The number of zeros added to right edge of the output.
-   * @param[in] invalid_bottom  The number of zeros added to top edge of the output.
-   * @param[in] weights_info    (Optional) Weights information needed for @ref CLConvolutionLayer,
-   *                            specifies if the weights tensor has been reshaped with @ref
-   *                            CLWeightsReshapeKernel.
    * @return a status
    */
   static Status validate(const ITensorInfo *input, const ITensorInfo *weights,
-                         const ITensorInfo *bias, ITensorInfo *output, const PadStrideInfo &info,
-                         unsigned int innvalid_right, unsigned int invalid_bottom,
+                         const ITensorInfo *bias, ITensorInfo *output,
+                         const PadStrideInfo &deconv_info, unsigned int invalid_right,
+                         unsigned int invalid_bottom,
                          const WeightsInfo &weights_info = WeightsInfo());
 
+  static DeconvolutionMethod
+  get_deconvolution_method(const ITensorInfo *input, const ITensorInfo *weights,
+                           const ITensorInfo *bias, ITensorInfo *output,
+                           const PadStrideInfo &deconv_info, unsigned int invalid_right,
+                           unsigned int invalid_bottom, const WeightsInfo &weights_info);
   // Inherited methods overridden:
   void run() override;
   void prepare() override;
 
 private:
-  MemoryGroup _memory_group;
-  CLTransposeConvLayerUpsample _scale_f;
-  CLConvolutionLayer _conv_f;
-  CPPFlipWeightsKernel _flip_weights;
-  CLTensor _scaled_output;
-  ICLTensor *_original_weights;
-  CLTensor _weights_flipped;
-  bool _is_prepared;
+  std::shared_ptr<IMemoryManager> _memory_manager;
+  std::unique_ptr<IFunction> _function;
 };
-}
+} // namespace arm_compute
 #endif /* __ARM_COMPUTE_CLTRANSPOSECONVLAYER_H__ */
diff --git a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLTransposeConvLayerUpsample.h b/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLTransposeConvLayerUpsample.h
deleted file mode 100644
index 7570fe76d..000000000
--- a/compute/ARMComputeEx/arm_compute/runtime/CL/functions/CLTransposeConvLayerUpsample.h
+++ /dev/null
@@ -1,102 +0,0 @@
-/*
- * Copyright (c) 2019 Samsung Electronics Co., Ltd. All Rights Reserved
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *      http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/*
- * Copyright (c) 2017-2018 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_CLTRANSPOSECONVLAYERUPSAMPLE_H__
-#define __ARM_COMPUTE_CLTRANSPOSECONVLAYERUPSAMPLE_H__
-
-#include "arm_compute/runtime/IFunction.h"
-
-#include "arm_compute/core/CL/kernels/CLTransposeConvLayerUpsampleKernel.h"
-#include "arm_compute/core/Types.h"
-#include "arm_compute/runtime/IFunction.h"
-#include "arm_compute/runtime/IMemoryManager.h"
-
-namespace arm_compute
-{
-class ICLTensor;
-
-/** Basic function to run @ref CLTransposeConvLayerUpsampleKernel */
-class CLTransposeConvLayerUpsample : public IFunction
-{
-public:
-  /** Default constructor */
-  CLTransposeConvLayerUpsample();
-  /** Prevent instances of this class from being copied (As this class contains pointers) */
-  CLTransposeConvLayerUpsample(const CLTransposeConvLayerUpsample &) = delete;
-  /** Prevent instances of this class from being copied (As this class contains pointers) */
-  CLTransposeConvLayerUpsample &operator=(const CLTransposeConvLayerUpsample &) = delete;
-  /** Allow instances of this class to be moved */
-  CLTransposeConvLayerUpsample(CLTransposeConvLayerUpsample &&) = default;
-  /** Allow instances of this class to be moved */
-  CLTransposeConvLayerUpsample &operator=(CLTransposeConvLayerUpsample &&) = default;
-  /** Default destructor */
-  virtual ~CLTransposeConvLayerUpsample() = default;
-
-  /** Initialize the function's source, destination, interpolation type and border_mode.
-   *
-   * @param[in, out] input        Source tensor. Data type supported: QASYMM8/F16/F32.
-   * @param[out]     output       Destination tensor. Data type supported: same as @p input.
-   * @param[in]      inner_border The number of zeros added to right and top edges of the input.
-   * @param[in]      info         Contains padding and policies to be used in the deconvolution.
-   */
-  void configure(ICLTensor *input, ICLTensor *output, const BorderSize &inner_border,
-                 const PadStrideInfo &info);
-  /** Static function to check if given info will lead to a valid configuration of @ref
-   * CLTransposeConvLayerUpsample
-   *
-   * @param[in] input        Source tensor info. Data type supported: QASYMM8/F16/F32.
-   * @param[in] output       Destination tensor info. Data type supported: same as @p input.
-   * @param[in] inner_border The number of zeros added to right and top edges of the input.
-   * @param[in] info         Contains padding and policies to be used in the deconvolution.
-   *
-   * @return a status
-   */
-  static Status validate(const ITensorInfo *input, const ITensorInfo *output,
-                         const BorderSize &inner_border, const PadStrideInfo &info);
-
-  // Inherited methods overridden:
-  void run() override;
-
-private:
-  CLTransposeConvLayerUpsampleKernel _upsample;
-  ICLTensor *_output;
-};
-}
-#endif /* __ARM_COMPUTE_CLTRANSPOSECONVLAYERUPSAMPLE_H__ */