diff options
| author | Michele Di Giorgio <michele.digiorgio@arm.com> | 2020-02-17 16:21:48 +0000 |
|---|---|---|
| committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2020-02-18 10:35:28 +0000 |
| commit | ca0ffa6243073140f2ba25d1830f0055e9f13d5f (patch) | |
| tree | 1952d4cd4eca62ebbacab23f6e90eed76c77e173 /src/core/NEON/kernels/NEPoolingLayerKernel.cpp | |
| parent | 2d22c3f9f130cd148b81caf79f2b21c9b4a89ad8 (diff) | |
| download | ComputeLibrary-ca0ffa6243073140f2ba25d1830f0055e9f13d5f.tar.gz | |
COMPMID-2819: Adding tolerance of 1 to CLColorConvert RGB8toU8 tests
Both rounding-to-zero and rounding-to-nearest even generate mismatches
with difference of 1 between reference and OpenCL implementation.
Therefore, this patch adds a tolerance to the test.
Change-Id: Id01a04c2b738ba0cc2981b675824cf3a838dc7e9
Signed-off-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/2737
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Georgios Pinitas <georgios.pinitas@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/NEPoolingLayerKernel.cpp')
0 files changed, 0 insertions, 0 deletions