diff options
author | Gian Marco <gianmarco.iodice@arm.com> | 2017-11-28 09:10:03 +0000 |
---|---|---|
committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:41:58 +0000 |
commit | 58c5794b917dae10ff115dd85ec69e2ca41136c1 (patch) | |
tree | f2cea2d94e6566be720256dc6105056798723699 /tests/validation/CPP/GEMMLowp.h | |
parent | 754e9526a7caf50876c2db9563dc72f096093b34 (diff) | |
download | ComputeLibrary-58c5794b917dae10ff115dd85ec69e2ca41136c1.tar.gz |
COMPMID-706 - Add GEMMLowp output stage for scaling by a fixed point number
DoD:
- Implement NEON kernel for quantizing down the gemmlowp result. The
result should be scaled by a fixedpoint number
- Implement OpenCL kernel for quantizing down the gemmlowp result. The
result should be scaled by a fixedpoint number
- Add test for validating the result
Required for:
- Integration of GEMMLowp in Android NN
- Convolution quantized
- Fully connected quantized
Change-Id: Ia963d25d695471e963961fb49a5600e78374ac4f
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/110981
Tested-by: BSG Visual Compute Jenkins server to access repositories on http://mpd-gerrit.cambridge.arm.com <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Diffstat (limited to 'tests/validation/CPP/GEMMLowp.h')
-rw-r--r-- | tests/validation/CPP/GEMMLowp.h | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/tests/validation/CPP/GEMMLowp.h b/tests/validation/CPP/GEMMLowp.h index 6c72b56e7a..a3d0bebe3f 100644 --- a/tests/validation/CPP/GEMMLowp.h +++ b/tests/validation/CPP/GEMMLowp.h @@ -49,6 +49,15 @@ SimpleTensor<T1> gemmlowp(const SimpleTensor<T2> &a, const SimpleTensor<T2> &b); template <typename T> SimpleTensor<uint8_t> gemmlowp_quantize_down_int32_to_uint8_scale(const SimpleTensor<T> &in, const SimpleTensor<T> &bias, int32_t result_offset, int32_t result_mult_int, int32_t result_shift, int32_t min = 0, int32_t max = 0); + +template <typename T> +SimpleTensor<uint8_t> gemmlowp_quantize_down_int32_to_uint8_scale_by_fixedpoint(const SimpleTensor<T> &in, int32_t result_fixedpoint_multiplier, int32_t result_shift, + int32_t result_offset_after_shift, + int32_t min = 0, int32_t max = 0); + +template <typename T> +SimpleTensor<uint8_t> gemmlowp_quantize_down_int32_to_uint8_scale_by_fixedpoint(const SimpleTensor<T> &in, const SimpleTensor<T> &bias, int32_t result_fixedpoint_multiplier, int32_t result_shift, + int32_t result_offset_after_shift, int32_t min = 0, int32_t max = 0); } // namespace reference } // namespace validation } // namespace test |