Add cropping support to NEBatchToSpace

- Deprecate dynamic block shape interface
- Iterate over output window instead of input window for simpler
  implementation and better performance
- Add cropping support and cropping tests

Resolves COMPMID-5918

Signed-off-by: SiCong Li <sicong.li@arm.com>
Change-Id: Ifea0f5f7760ffd0f4d5d4f3a5ae8d14d0b98b790
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9378
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 13 insertions, 6 deletions

diff --git a/src/core/NEON/kernels/NEBatchToSpaceLayerKernel.h b/src/core/NEON/kernels/NEBatchToSpaceLayerKernel.h
index 26e8224922..5eceee0904 100644
--- a/src/core/NEON/kernels/NEBatchToSpaceLayerKernel.h
+++ b/src/core/NEON/kernels/NEBatchToSpaceLayerKernel.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2019-2020 Arm Limited.
+ * Copyright (c) 2019-2020, 2023 Arm Limited.
  *
  * SPDX-License-Identifier: MIT
  *
@@ -24,6 +24,7 @@
 #ifndef ARM_COMPUTE_NEBATCHTOSPACELAYERKERNEL_H
 #define ARM_COMPUTE_NEBATCHTOSPACELAYERKERNEL_H
 
+#include "arm_compute/core/Types.h"
 #include "src/core/NEON/INEKernel.h"
 
 namespace arm_compute
@@ -55,6 +56,8 @@ public:
      * @param[in]  input       Tensor input. Supported tensor rank: 4. Data types supported: All.
      * @param[in]  block_shape 1-D tensor with shape [M]. Data types supported: S32
      * @param[out] output      Tensor output. Data types supported: same as @p input
+     *
+     * @deprecated This method for dynamic block shape is not fully mature and will be removed in 23.08 release
      */
     void configure(const ITensor *input, const ITensor *block_shape, ITensor *output);
     /** Initialise the kernel's inputs and output (Static block shape).
@@ -63,8 +66,9 @@ public:
      * @param[in]  block_shape_x Block shape x value.
      * @param[in]  block_shape_y Block shape y value.
      * @param[out] output        Tensor output. Data types supported: same as @p input
+     * @param[in]  crop_info     Specifies how the output shape is cropped after batch to space is performed
      */
-    void configure(const ITensor *input, const int32_t block_shape_x, const int32_t block_shape_y, ITensor *output);
+    void configure(const ITensor *input, int32_t block_shape_x, int32_t block_shape_y, ITensor *output, const CropInfo &crop_info = CropInfo{});
     /** Static function to check if given info will lead to a valid configuration of @ref NEBatchToSpaceLayerKernel
      *
      * @param[in] input       Tensor input. Supported tensor rank: 4. Data types supported: All.
@@ -72,6 +76,8 @@ public:
      * @param[in] output      Tensor output. Data types supported: same as @p input
      *
      * @return a status
+     *
+     * @deprecated This method for dynamic block shape is not fully mature and will be removed in 23.08 release
      */
     static Status validate(const ITensorInfo *input, const ITensorInfo *block_shape, const ITensorInfo *output);
     /** Static function to check if given info will lead to a valid configuration of @ref NEBatchToSpaceLayerKernel (Static block shape).
@@ -80,10 +86,11 @@ public:
      * @param[in] block_shape_x Block shape x value.
      * @param[in] block_shape_y Block shape y value.
      * @param[in] output        Tensor output. Data types supported: same as @p input
+     * @param[in] crop_info     Specifies how the output shape is cropped after batch to space is performed
      *
      * @return a status
      */
-    static Status validate(const ITensorInfo *input, const int32_t block_shape_x, const int32_t block_shape_y, const ITensorInfo *output);
+    static Status validate(const ITensorInfo *input, int32_t block_shape_x, int32_t block_shape_y, const ITensorInfo *output, const CropInfo &crop_info = CropInfo{});
 
     // Inherited methods overridden:
     void run(const Window &window, const ThreadInfo &info) override;
@@ -93,9 +100,9 @@ private:
     const ITensor *_block_shape; /**< Block shape tensor */
     ITensor       *_output;      /**< Destination tensor */
     DataLayout     _data_layout; /**< Data layout to  be used at run-time */
-
-    int32_t _block_shape_x;
-    int32_t _block_shape_y;
+    int32_t        _block_shape_x;
+    int32_t        _block_shape_y;
+    CropInfo       _crop_info; /**< Information related to cropping performed on output after the operation */
 };
 } // namespace arm_compute
 #endif /* ARM_COMPUTE_NEBATCHTOSPACELAYERKERNEL_H */