aboutsummaryrefslogtreecommitdiff
path: root/arm_compute/core/NEON/kernels/NEBatchNormalizationLayerKernel.h
diff options
context:
space:
mode:
authorMichele Di Giorgio <michele.digiorgio@arm.com>2018-03-01 16:56:48 +0000
committerAnthony Barbier <anthony.barbier@arm.com>2018-11-02 16:49:54 +0000
commit0cbb927ac309e332ac6e6f1ab9170f041f0138ab (patch)
tree102d50dec9f741f04b1126ae03e6e491dda2d3ba /arm_compute/core/NEON/kernels/NEBatchNormalizationLayerKernel.h
parent82b51482479951cf133c223eb81aae291cb4d590 (diff)
downloadComputeLibrary-0cbb927ac309e332ac6e6f1ab9170f041f0138ab.tar.gz
COMPMID-804: Add NHWC data format support for NEON batch normalisation
Change-Id: I04892e7be3f5aa58cd95917a4f90a6b4ffcf6efc Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/122897 Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Tested-by: Jenkins <bsgcomp@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Diffstat (limited to 'arm_compute/core/NEON/kernels/NEBatchNormalizationLayerKernel.h')
-rw-r--r--arm_compute/core/NEON/kernels/NEBatchNormalizationLayerKernel.h21
1 files changed, 19 insertions, 2 deletions
diff --git a/arm_compute/core/NEON/kernels/NEBatchNormalizationLayerKernel.h b/arm_compute/core/NEON/kernels/NEBatchNormalizationLayerKernel.h
index ae6b8634b3..2d33f87dfa 100644
--- a/arm_compute/core/NEON/kernels/NEBatchNormalizationLayerKernel.h
+++ b/arm_compute/core/NEON/kernels/NEBatchNormalizationLayerKernel.h
@@ -119,7 +119,15 @@ private:
* @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()).
*/
template <bool fused_activation>
- void batch_normalization_fp16(const Window &window);
+ void batch_normalization_fp16_nchw(const Window &window);
+ /** Template function to run batch normalization on fp16 on tensors with NHWC format
+ *
+ * @tparam fused_activation Boolean that flags if its a fused activation or not
+ *
+ * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()).
+ */
+ template <bool fused_activation>
+ void batch_normalization_fp16_nhwc(const Window &window);
/** Template function to run batch normalization on fp32
*
* @tparam fused_activation Boolean that flags if its a fused activation or not
@@ -128,7 +136,16 @@ private:
* @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()).
*/
template <bool fused_activation, typename F>
- void batch_normalization_fp32(const Window &window);
+ void batch_normalization_fp32_nchw(const Window &window);
+ /** Template function to run batch normalization on fp32 on tensors with NHWC format
+ *
+ * @tparam fused_activation Boolean that flags if its a fused activation or not
+ * @tparam F Activation function functor to run
+ *
+ * @param[in] window Region on which to execute the kernel. (Must be a valid region of the window returned by window()).
+ */
+ template <bool fused_activation, typename F>
+ void batch_normalization_fp32_nhwc(const Window &window);
/** Common signature for all the batch normalization functions
*
* @param[in] window Region on which to execute the kernel.