COMPMID-3576: Nightly failure: NEON/PoolingLayer/Float/FP16/MaxUnpooling S10
Extend NEPoolingLayer max pooling to extract indices for FP16
Signed-off-by: Sheri Zhang <sheri.zhang@arm.com>
Change-Id: I5a7c754be353e4c2c5d0ab3794e9427408d0c4fa
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/3580
Reviewed-by: Sang-Hoon Park <sang-hoon.park@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
diff --git a/arm_compute/core/NEON/kernels/NEMaxUnpoolingLayerKernel.h b/arm_compute/core/NEON/kernels/NEMaxUnpoolingLayerKernel.h
index 9c9b194..f3ea049 100644
--- a/arm_compute/core/NEON/kernels/NEMaxUnpoolingLayerKernel.h
+++ b/arm_compute/core/NEON/kernels/NEMaxUnpoolingLayerKernel.h
@@ -55,7 +55,10 @@
* @note Output shape must be equal to the shape of the original input to pool.
*
* @param[in] input Source tensor. Data types supported: QASYMM8/QASYMM8_SIGNED/F16/F32.
- * @param[out] indices The indices of the maximal values. Data type supported: U32.
+ * @param[in] indices Tensor containing the offset to store the input elements in the output tensor.
+ * @ref NEPoolingLayerKernel with indices should precede this function in order to
+ * properly reconstruct the output tensor.
+ * The tensor shape of this tensor has to be equal to the input tensor shape. Data type supported: U32.
* @param[out] output Destination tensor. Data types supported: Same as @p input.
* @param[in] pool_info Contains pooling operation information described in @ref PoolingLayerInfo.
*/