diff options
| author | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2019-06-10 14:46:49 +0100 |
|---|---|---|
| committer | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2019-06-11 12:08:08 +0000 |
| commit | 761c8d02ff875877db7aa7c850cf8d128592e822 (patch) | |
| tree | 10871f3dccfa262d4a051d3d88b899be6acac0a2 /arm_compute/core/CL | |
| parent | d5134364fc4ca40ea65635192e7959327d690a01 (diff) | |
| download | ComputeLibrary-761c8d02ff875877db7aa7c850cf8d128592e822.tar.gz | |
COMPMID-2398: Add test for CLFuseBatchNormalizationLayer
Change-Id: I786df628ce15fc33fc42c9437fe82972e02e3b16
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/1317
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'arm_compute/core/CL')
| -rw-r--r-- | arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h | 32 |
1 files changed, 18 insertions, 14 deletions
diff --git a/arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h b/arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h index 05a57c171e..a5b98bb27d 100644 --- a/arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h +++ b/arm_compute/core/CL/kernels/CLFuseBatchNormalizationKernel.h @@ -52,11 +52,13 @@ public: * @param[in] conv_weights Convolution layer weights tensor. Data type supported: F16/F32 * @param[in] bn_mean Batch normalization layer mean tensor. Same as @p conv_weights * @param[in] bn_var Batch normalization layer variance tensor. Same as @p conv_weights - * @param[out] fused_weights Output fused weights tensor. Same as @p conv_weights - * @param[out] fused_bias Output fused bias tensor. Same as @p conv_weights - * @param[in] conv_bias (Optional) Convolution layer bias tensor. Same as @p conv_weights - * @param[in] bn_beta (Optional) Batch normalization layer beta tensor. Same as @p conv_weights - * @param[in] bn_gamma (Optional) Batch normalization layer gamma tensor. Same as @p conv_weights + * @param[out] fused_weights Output fused weights tensor. It can be a nullptr in case of in-place computation. Same as @p conv_weights + * @param[out] fused_bias Output fused bias tensor. It can be a nullptr in case of in-place computation and conv_bias != nullptr. Same as @p conv_weights + * @param[in] conv_bias (Optional) Convolution layer bias tensor. It can be a nullptr in case the bias tensor is not required. Same as @p conv_weights + * @param[in] bn_beta (Optional) Batch normalization layer beta tensor. It can be a nullptr in case the beta tensor is not required. Same as @p conv_weights + * @note if nullptr, bn_beta is set to 0.0 + * @param[in] bn_gamma (Optional) Batch normalization layer gamma tensor. It can be a nullptr in case the gamma tensor is not required. Same as @p conv_weights + * @note if nullptr, bn_gamma is set to 1.0 * @param[in] epsilon (Optional) Batch normalization layer epsilon parameter. Defaults to 0.001f. */ void configure(const ICLTensor *conv_weights, const ICLTensor *bn_mean, const ICLTensor *bn_var, ICLTensor *fused_weights, ICLTensor *fused_bias, @@ -64,15 +66,17 @@ public: float epsilon = 0.001f); /** Static function to check if given info will lead to a valid configuration of @ref CLFuseBatchNormalizationKernel * - * @param[in] conv_weights Convolution layer weights tensor. Data type supported: F16/F32 - * @param[in] bn_mean Batch normalization layer mean tensor. Same as @p conv_weights - * @param[in] bn_var Batch normalization layer variance tensor. Same as @p conv_weights - * @param[in] fused_weights Output fused weights tensor. Same as @p conv_weights - * @param[in] fused_bias Output fused bias tensor. Same as @p conv_weights - * @param[in] conv_bias (Optional) Convolution layer bias tensor. Same as @p conv_weights - * @param[in] bn_beta (Optional) Batch normalization layer beta tensor. Same as @p conv_weights - * @param[in] bn_gamma (Optional) Batch normalization layer gamma tensor. Same as @p conv_weights - * @param[in] epsilon (Optional) Batch normalization layer epsilon parameter. Defaults to 0.001f. + * @param[in] conv_weights Convolution layer weights tensor info. Data type supported: F16/F32 + * @param[in] bn_mean Batch normalization layer mean tensor info. Same as @p conv_weights + * @param[in] bn_var Batch normalization layer variance tensor info. Same as @p conv_weights + * @param[out] fused_weights Output fused weights tensor info. It can be a nullptr in case of in-place computation. Same as @p conv_weights + * @param[out] fused_bias Output fused bias tensor info. It can be a nullptr in case of in-place computation and conv_bias != nullptr. Same as @p conv_weights + * @param[in] conv_bias (Optional) Convolution layer bias tensor info. It can be a nullptr in case the bias tensor is not required. Same as @p conv_weights + * @param[in] bn_beta (Optional) Batch normalization layer beta tensor info. It can be a nullptr in case the beta tensor is not required. Same as @p conv_weights + * @note if nullptr, bn_beta is set to 0.0 + * @param[in] bn_gamma (Optional) Batch normalization layer gamma tensor info. It can be a nullptr in case the gamma tensor is not required. Same as @p conv_weights + * @note if nullptr, bn_gamma is set to 1.0 + * @param[in] epsilon (Optional) Batch normalization layer epsilon parameter. Defaults to 0.001f. * * @return a status */ |