diff options
author | Michele Di Giorgio <michele.digiorgio@arm.com> | 2019-10-29 10:58:13 +0000 |
---|---|---|
committer | Michele Di Giorgio <michele.digiorgio@arm.com> | 2019-12-20 14:05:24 +0000 |
commit | f29d1b7d8bf2d1619554eb3443556b44d4aa1a4c (patch) | |
tree | 0a427f7fda2131f39e055f27b97f0a612aff990c /arm_compute/runtime/NEON/functions | |
parent | 748a7c81245ae81d04607b3a762cf65cd39026f2 (diff) | |
download | ComputeLibrary-f29d1b7d8bf2d1619554eb3443556b44d4aa1a4c.tar.gz |
COMPMID-2608: Enable quantization with multiplier greater than 1 on NEON
Change-Id: Ib2b0c9ac88fc2b645f478c9981f71ee28f2c77fd
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/2425
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'arm_compute/runtime/NEON/functions')
-rw-r--r-- | arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h | 24 | ||||
-rw-r--r-- | arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h | 8 |
2 files changed, 17 insertions, 15 deletions
diff --git a/arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h b/arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h index 8150737ebe..784637a796 100644 --- a/arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h +++ b/arm_compute/runtime/NEON/functions/NEFullyConnectedLayer.h @@ -131,7 +131,7 @@ public: * If this function is called after a Convolution Layer, the (transposed) weights will have as many rows as the product of the first 3 input's dimensions. * If it is called after another FullyConnected Layer, the (transposed) weights will have as many rows as the input's first dimension. * Data type supported: Same as @p input. - * @param[in] biases Bias tensor. Can be nullptr. Data type supported:Same as @p input. + * @param[in] biases Bias tensor. Can be nullptr. Data type supported: Same as @p weights, S32 if @p weights is QASYMM8. * @param[out] output Destination tensor. Its shape should be equal to the output of a matrix multiplication between: * - The output of im2col on the input and the (transposed) 2D weights, if the function is called after a Convolution Layer * - The input tensor and the (transposed) 2D weights, if the function is called after another FullyConnected Layer. @@ -142,17 +142,17 @@ public: FullyConnectedLayerInfo fc_info = FullyConnectedLayerInfo()); /** Static function to check if given info will lead to a valid configuration of @ref NEFullyConnectedLayer * - * @param[in] input Source tensor info. Data type supported: QASYMM8/F16/F32. - * @param[in] weights Weights tensor info. The weights must be 2 dimensional. - * If this function is called after a Convolution Layer, the (transposed) weights will have as many rows as the product of the first 3 input's dimensions. - * If it is called after another FullyConnected Layer, the (transposed) weights will have as many rows as the input's first dimension. - * Data type supported: Same as @p input. - * @param[in] biases Bias tensor info. Can be nullptr. Data type supported:Same as @p input. - * @param[out] output Destination tensor info. Its shape should be equal to the output of a matrix multiplication between: - * - The output of im2col on the input and the (transposed) 2D weights, if the function is called after a Convolution Layer - * - The input tensor and the (transposed) 2D weights, if the function is called after another FullyConnected Layer. - * Data type supported: Same as @p input. - * @param[in] fc_info (Optional) Fully connected layer additional info + * @param[in] input Source tensor info. Data type supported: QASYMM8/F16/F32. + * @param[in] weights Weights tensor info. The weights must be 2 dimensional. + * If this function is called after a Convolution Layer, the (transposed) weights will have as many rows as the product of the first 3 input's dimensions. + * If it is called after another FullyConnected Layer, the (transposed) weights will have as many rows as the input's first dimension. + * Data type supported: Same as @p input. + * @param[in] biases Bias tensor. Can be nullptr. Data type supported: Same as @p weights, S32 if @p weights is QASYMM8. + * @param[in] output Destination tensor info. Its shape should be equal to the output of a matrix multiplication between: + * - The output of im2col on the input and the (transposed) 2D weights, if the function is called after a Convolution Layer + * - The input tensor and the (transposed) 2D weights, if the function is called after another FullyConnected Layer. + * Data type supported: Same as @p input. + * @param[in] fc_info (Optional) Fully connected layer additional info * * @return a status */ diff --git a/arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h b/arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h index 665d4f1bae..660f55953e 100644 --- a/arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h +++ b/arm_compute/runtime/NEON/functions/NEGEMMConvolutionLayer.h @@ -64,15 +64,17 @@ public: /** Set the input and output tensors. * * @param[in] weights Weights tensor. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM]. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL/F16/F32. - * @param[in] biases Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p weights. + * @param[in] biases Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM]. + * Data type supported: Same as @p weights, S32 if @p weights is QASYMM8/QASYMM8_SIGNED. * @param[out] output Destination tensor. Data types supported: Same as @p weights. */ void configure(const ITensor *weights, const ITensor *biases, ITensor *output); /** Static function to check if given info will lead to a valid configuration of @ref NEConvolutionLayerReshapeWeights * * @param[in] weights Weights tensor info. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM]. Data type supported: QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL/F16/F32. - * @param[in] biases Biases tensor info. Shared biases supported. Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as @p weights. - * @param[in] output Destination tensor info. Data types supported: Same as @p weights. + * @param[in] biases Biases tensor. Shared biases supported. Biases are 1D tensor with dimensions [OFM]. + * Data type supported: Same as @p weights, S32 if @p weights is QASYMM8/QASYMM8_SIGNED. + * @param[in] output Destination tensor. Data types supported: Same as @p weights. * * @return an error status */ |