From 2213d4b334567d0cb7f283090d42b5fb1b70f66b Mon Sep 17 00:00:00 2001 From: Gian Marco Iodice Date: Fri, 27 Apr 2018 10:39:06 +0100 Subject: COMPMID-1096 - Add fast_math flag to CLConvolutionLayer COMPMID-1103 - CLWinogradConvolutionLayer mismatches Change-Id: Iceaa9482a1790ec39d2720c220261aaea8043978 Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/129398 Tested-by: Jenkins Reviewed-by: Giorgio Arena Reviewed-by: Georgios Pinitas --- arm_compute/runtime/NEON/functions/NEConvolutionLayer.h | 2 +- arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h | 10 +++++----- arm_compute/runtime/NEON/functions/NEReshapeLayer.h | 4 ++-- 3 files changed, 8 insertions(+), 8 deletions(-) (limited to 'arm_compute/runtime/NEON') diff --git a/arm_compute/runtime/NEON/functions/NEConvolutionLayer.h b/arm_compute/runtime/NEON/functions/NEConvolutionLayer.h index ce9a3ed4f2..b82ba89f7c 100644 --- a/arm_compute/runtime/NEON/functions/NEConvolutionLayer.h +++ b/arm_compute/runtime/NEON/functions/NEConvolutionLayer.h @@ -52,7 +52,7 @@ public: * * @param[in] input Source tensor. 3 lower dimensions represent a single input [width, height, IFM], * while every optional dimension from 4 and above represent a batch of inputs. - * Data types supported: QS8/QASYMM8/QS16/F32. + * Data types supported: QS8/QASYMM8/QS16/F16/F32. * @param[in] weights Weights tensor. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM]. Data type supported: Same as @p input. * @param[in] biases Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM]. * Data type supported: Should match @p input data type, except for input of QASYMM8 type where biases should be of S32 type. diff --git a/arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h b/arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h index 7d5e397e80..ba96ae6cfa 100644 --- a/arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h +++ b/arm_compute/runtime/NEON/functions/NEPixelWiseMultiplication.h @@ -37,11 +37,11 @@ class NEPixelWiseMultiplication : public INESimpleFunction public: /** Initialise the kernel's inputs, output and convertion policy. * - * @param[in, out] input1 An input tensor. Data types supported: U8/QS8/S16/F32. + * @param[in, out] input1 An input tensor. Data types supported: U8/QS8/S16/F16/F32. * The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0. * @param[in, out] input2 An input tensor. Data types supported: same as @p input1. * The input tensor is [in, out] because its TensorInfo might be modified inside the kernel in case of broadcasting of dimension 0. - * @param[out] output Output tensor. Data types supported: U8/QS8/S16/F32. + * @param[out] output Output tensor. Data types supported: U8/QS8/S16/F16/F32. * @param[in] scale Scale to apply after multiplication. * Scale must be positive and its value must be either 1/255 or 1/2^n where n is between 0 and 15. For QS8 and QS16 scale must be 1. * @param[in] overflow_policy Overflow policy. @@ -50,9 +50,9 @@ public: void configure(ITensor *input1, ITensor *input2, ITensor *output, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy); /** Static function to check if given info will lead to a valid configuration of @ref NEPixelWiseMultiplication * - * @param[in] input1 First tensor info input. Data types supported: U8/QS8/S16/F32. - * @param[in] input2 Second tensor info input. Data types supported: U8/QS8/S16/F32. - * @param[in] output Output tensor info. Data types supported: U8/QS8/S16/F32. + * @param[in] input1 First tensor info input. Data types supported: U8/QS8/S16/F16/F32. + * @param[in] input2 Second tensor info input. Data types supported: U8/QS8/S16/F16/F32. + * @param[in] output Output tensor info. Data types supported: U8/QS8/S16/F16/F32. * @param[in] scale Scale to apply after multiplication. Must be positive. * @param[in] overflow_policy Overflow policy. * @param[in] rounding_policy Rounding policy. diff --git a/arm_compute/runtime/NEON/functions/NEReshapeLayer.h b/arm_compute/runtime/NEON/functions/NEReshapeLayer.h index 369f50e147..0bab534ebc 100644 --- a/arm_compute/runtime/NEON/functions/NEReshapeLayer.h +++ b/arm_compute/runtime/NEON/functions/NEReshapeLayer.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2017 ARM Limited. + * Copyright (c) 2017-2018 ARM Limited. * * SPDX-License-Identifier: MIT * @@ -37,7 +37,7 @@ class NEReshapeLayer : public INESimpleFunction public: /** Initialise the kernel's inputs and outputs * - * @param[in] input First tensor input. Data type supported: U8/S8/QS8/U16/S16/QS16/U32/S32/F16/F32 + * @param[in] input First tensor input. Data type supported: U8/S8/QS8/QASYMM8//U16/S16/QS16/U32/S32/F16/F32 * @param[out] output Output tensor. Data type supported: Same as @p input */ void configure(const ITensor *input, ITensor *output); -- cgit v1.2.1