diff options
author | Georgios Pinitas <georgios.pinitas@arm.com> | 2020-02-26 09:58:13 +0000 |
---|---|---|
committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2020-03-05 15:15:15 +0000 |
commit | e8291acc1d9e89c9274d31f0d5bb4779eb95588c (patch) | |
tree | 5a0fef36d6daabe387174e55b60de54557c75291 /arm_compute/runtime | |
parent | aa85cdf22802cb892d7fa422ca505a43d84adb38 (diff) | |
download | ComputeLibrary-e8291acc1d9e89c9274d31f0d5bb4779eb95588c.tar.gz |
COMPMID-3152: Initial Bfloat16 support
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: Ie6959e37e13731c86b2ee29392a99a293450a1b4
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/2824
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Diffstat (limited to 'arm_compute/runtime')
-rw-r--r-- | arm_compute/runtime/NEON/functions/NEDepthConvertLayer.h | 23 |
1 files changed, 12 insertions, 11 deletions
diff --git a/arm_compute/runtime/NEON/functions/NEDepthConvertLayer.h b/arm_compute/runtime/NEON/functions/NEDepthConvertLayer.h index 43a256ebe2..b784480887 100644 --- a/arm_compute/runtime/NEON/functions/NEDepthConvertLayer.h +++ b/arm_compute/runtime/NEON/functions/NEDepthConvertLayer.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2016-2019 ARM Limited. + * Copyright (c) 2016-2020 ARM Limited. * * SPDX-License-Identifier: MIT * @@ -47,23 +47,24 @@ public: * * Valid conversions Input -> Output : * - * - QASYMM8 -> F16, F32 - * - U8 -> U16, S16, S32 - * - U16 -> U8, U32 - * - S16 -> U8, S32 - * - F16 -> QASYMM8, F32 - * - F32 -> QASYMM8, F16 + * - QASYMM8 -> F16, F32 + * - U8 -> U16, S16, S32 + * - U16 -> U8, U32 + * - S16 -> U8, S32 + * - BFLOAT16 -> F32 + * - F16 -> QASYMM8, F32 + * - F32 -> QASYMM8, F16, BFLOAT16 * - * @param[in] input The input tensor to convert. Data types supported: QASYMM8/U8/U16/S16/F16/F32. - * @param[out] output The output tensor. Data types supported: QASYMM8/U8/U16/S16/U32/S32/F16/F32. + * @param[in] input The input tensor to convert. Data types supported: QASYMM8/U8/U16/S16/BFLOAT16/F16/F32. + * @param[out] output The output tensor. Data types supported: QASYMM8/U8/U16/S16/U32/S32/BFLOAT16/F16/F32. * @param[in] policy Conversion policy. * @param[in] shift (Optional) Value for down/up conversions. Must be 0 <= shift < 8. */ void configure(const ITensor *input, ITensor *output, ConvertPolicy policy, uint32_t shift = 0); /** Static function to check if given info will lead to a valid configuration of @ref NEDepthConvertLayer * - * @param[in] input Source tensor info. Data types supported: QASYMM8/U8/U16/S16/F16/F32. - * @param[in] output Destination tensor info. Data type supported: QASYMM8/U8/U16/S16/U32/S32/F16/F32. + * @param[in] input Source tensor info. Data types supported: QASYMM8/U8/U16/S16/BFLOAT16/F16/F32. + * @param[in] output Destination tensor info. Data type supported: QASYMM8/U8/U16/S16/U32/S32/BFLOAT16/F16/F32. * @param[in] policy Conversion policy. * @param[in] shift (Optional) Value for down/up conversions. Must be 0 <= shift < 8. * |