aboutsummaryrefslogtreecommitdiff
path: root/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
diff options
context:
space:
mode:
authorGeorgios Pinitas <georgios.pinitas@arm.com>2017-11-17 10:55:00 +0000
committerAnthony Barbier <anthony.barbier@arm.com>2018-11-02 16:35:24 +0000
commit540d008180a39a84770e51b4bc891cd2ff85980e (patch)
tree500faadfafde7b92bdf7cb46c01501d2db01728f /arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
parent43ce898014faabbe8ab1cdf714b7aad3d3c9b2a9 (diff)
downloadComputeLibrary-540d008180a39a84770e51b4bc891cd2ff85980e.tar.gz
COMPMID-556: Fixes bias in CLDirectConvolutionLayer to be int32.
Biases were incorrectly passed as uchar8 in asymmetric quantized calculations while they should be in int32. Change-Id: I461f5e4ef6eb44374c1094378a1bfe9ade5e247d Reviewed-on: http://mpd-gerrit.cambridge.arm.com/96244 Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Diffstat (limited to 'arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h')
-rw-r--r--arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h3
1 files changed, 2 insertions, 1 deletions
diff --git a/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h b/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
index 85deeaef37..da786d71df 100644
--- a/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
+++ b/arm_compute/core/CL/kernels/CLDirectConvolutionLayerKernel.h
@@ -60,7 +60,8 @@ public:
* @param[in] weights Weights tensor. Weights are 4D tensor with dimensions [kernel_x, kernel_y, IFM, OFM].
* The 3rd dimension must be the same as the input's volume 3rd dimension.
* Data type supported:Same as @p input.
- * @param[in] biases Biases tensor. Biases are 1D tensor with dimension [OFM]. Data type supported: Same as @p input.
+ * @param[in] biases Biases tensor. Biases are 1D tensor with dimension [OFM].
+ * Data type supported: Should match @p input data type, except for input of QASYMM8 type where biases should be of S32 type
* @param[out] output Output tensor.
* The 3rd dimensions must be equal to the 4th dimension of the @p kernels tensor. Data types supported: Same as @p input.
* @param[in] conv_info Contains padding and stride information described in @ref PadStrideInfo.