diff options
author | Chunosov <N.Chunosov@yandex.ru> | 2017-11-06 22:09:45 +0700 |
---|---|---|
committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:35:24 +0000 |
commit | d6afedc775220f17317f1835a4d18b72a54525de (patch) | |
tree | 54aed8322a4a286ba376d74bbee61c85a588cc9b /arm_compute/runtime/CL/functions | |
parent | 6ff12a0f7765f62b8d0fa8554021e1cac2789f19 (diff) | |
download | ComputeLibrary-d6afedc775220f17317f1835a4d18b72a54525de.tar.gz |
COMPMID-661: softmax-fp32 optimisation (#14)
Change-Id: I2007af1ed9dcf68065cf412aa50f73a2025b31a6
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/94605
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Diffstat (limited to 'arm_compute/runtime/CL/functions')
-rw-r--r-- | arm_compute/runtime/CL/functions/CLSoftmaxLayer.h | 18 |
1 files changed, 10 insertions, 8 deletions
diff --git a/arm_compute/runtime/CL/functions/CLSoftmaxLayer.h b/arm_compute/runtime/CL/functions/CLSoftmaxLayer.h index d84297e9a1..72ef679d6a 100644 --- a/arm_compute/runtime/CL/functions/CLSoftmaxLayer.h +++ b/arm_compute/runtime/CL/functions/CLSoftmaxLayer.h @@ -54,8 +54,8 @@ public: /** Set the input and output tensors. * * @param[in] input Source tensor. Data types supported: QS8/QS16/F16/F32 - * @param[in] beta A scaling factor for the exponent. * @param[out] output Destination tensor. Data types supported: same as @p input + * @param[in] beta (Optional) A scaling factor for the exponent. Defaults to 1.f */ void configure(const ICLTensor *input, ICLTensor *output, float beta = 1.0f); @@ -63,13 +63,15 @@ public: void run() override; private: - CLMemoryGroup _memory_group; - CLLogits1DMaxKernel _max_kernel; - CLLogits1DShiftExpSumKernel _shift_exp_sum_kernel; - CLLogits1DNormKernel _norm_kernel; - CLTensor _max; - CLTensor _sum; - CLTensor _tmp; + CLMemoryGroup _memory_group; + CLLogits1DMaxKernel _max_kernel; + CLLogits1DShiftExpSumKernel _shift_exp_sum_kernel; + CLLogits1DMaxShiftExpSumKernel _max_shift_exp_sum_kernel; + CLLogits1DNormKernel _norm_kernel; + CLTensor _max; + CLTensor _sum; + CLTensor _tmp; + bool _run_legacy_path; }; } #endif /* __ARM_COMPUTE_CLSOFTMAXLAYER_H__ */ |