diff options
| author | Chunosov <N.Chunosov@yandex.ru> | 2017-11-08 16:09:35 +0700 |
|---|---|---|
| committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:35:24 +0000 |
| commit | f450caa7d2ac9a2a90407fb81203228dc82ef4a1 (patch) | |
| tree | ed34d43943cd36cbd6776ddc6ac87e92d6f7dcc0 /arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h | |
| parent | 7068f9900d136312318ff430aef588b14e0c87ad (diff) | |
| download | ComputeLibrary-f450caa7d2ac9a2a90407fb81203228dc82ef4a1.tar.gz | |
COMPMID-661: softmax-uint8 implementation (#16)
Change-Id: Iad11ce70a8a0878a48e445a092035c49c926cece
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/94855
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Diffstat (limited to 'arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h')
| -rw-r--r-- | arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h | 17 |
1 files changed, 9 insertions, 8 deletions
diff --git a/arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h b/arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h index 675c462c95..ab550aa52a 100644 --- a/arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h +++ b/arm_compute/core/CL/kernels/CLSoftmaxLayerKernel.h @@ -38,7 +38,7 @@ class CLLogits1DMaxKernel : public ICLSimple3DKernel public: /** Set the input and output tensors. * - * @param[in] input Source tensor. Data types supported: QS8/QS16/F16/F32 + * @param[in] input Source tensor. Data types supported: QS8/QASYMM8/QS16/F16/F32 * @param[out] output Destination tensor. Data types supported: same as @p input */ void configure(const ICLTensor *input, ICLTensor *output); @@ -60,11 +60,11 @@ public: CLLogits1DShiftExpSumKernel &operator=(CLLogits1DShiftExpSumKernel &&) = default; /** Set the input and output tensors. * - * @param[in] input Source tensor. Data types supported: QS8/QS16/F16/F32 + * @param[in] input Source tensor. Data types supported: QS8/QASYMM8/QS16/F16/F32 * @param[in] max Max values tensor. Data types supported: same as @p input - * @param[out] output Destination tensor. Data types supported: same as @p input - * @param[out] sum Sum of 1D logits tensor. Data types supported: same as @p input - * @param[in] beta (Optional) A scaling factor for the exponent. Defaults to 1.f + * @param[out] output Destination tensor. Data types supported: S32 for QASYMM8 @p input, or same as @p input + * @param[out] sum Sum of 1D logits tensor. Data types supported: S32 for QASYMM8 @p input, or same as @p input + * @param[in] beta (Optional) A scaling factor for the exponent. Defaults to 1.0 */ void configure(const ICLTensor *input, const ICLTensor *max, ICLTensor *output, ICLTensor *sum, float beta = 1.0f); @@ -146,11 +146,12 @@ public: CLLogits1DNormKernel &operator=(CLLogits1DNormKernel &&) = default; /** Set the input and output tensors. * - * @param[in] input Source tensor. Data types supported: QS8/QS16/F16/F32 + * @param[in] input Source tensor. Data types supported: QS8/QS16/S32/F16/F32 * @param[in] sum Sum tensor. Dimensions should be dim(input)-1. Data types supported: same as @p input - * @param[out] output Destination tensor. Data types supported: same as @p input + * @param[out] output Destination tensor. Data types supported: QASYMM8 for S32 @p input, or same as @p input + * @param[in] beta (Optional) A scaling factor for the exponent. (Default = 1.0) */ - void configure(const ICLTensor *input, const ICLTensor *sum, ICLTensor *output); + void configure(const ICLTensor *input, const ICLTensor *sum, ICLTensor *output, float beta = 1.0f); // Inherited methods overridden: void run(const Window &window, cl::CommandQueue &queue) override; |