tequilaOS/platform_hardware_interfaces
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Add TENSOR_QUANT8_ASYMM_SIGNED support for pooling ops

Browse source 
* Update AVERAGE_POOL_2D and MAX_POOL_2D

Bug: 143934716
Bug: 143934303
Test: quantization coupling tests in CTS and VTS
Change-Id: I15f36d6f819c92f96a46c40f9cc1491627644278
Merged-In: I15f36d6f819c92f96a46c40f9cc1491627644278
(cherry picked from commit 3f17cc3daa)
This commit is contained in:
Lev Proleev 2019-11-20 13:55:16 +00:00
parent 53c79d179a
commit d1b3450b3d
2 changed files with 7 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
current.txt
View file

@ -597,7 +597,7 @@ adb0efdf1462e9b2e742c0dcadd598666aac551f178be06e755bfcdf5797abd0 android.hardwar
9e59fffceed0dd72a9799e04505db5f777bbbea1af0695ba4107ef6d967c6fda android.hardware.neuralnetworks@1.3::IDevice
4a6c3b3556da951b4def21ba579a227c022980fe4465df6cdfbe20628fa75f5a android.hardware.neuralnetworks@1.3::IPreparedModel
94e803236398bed1febb11cc21051bc42ec003700139b099d6c479e02a7ca3c3 android.hardware.neuralnetworks@1.3::IPreparedModelCallback
103cb87c5ed46851badac097f8d190da60f39b5ab32d60e2e93f64d3014ea75c android.hardware.neuralnetworks@1.3::types
6256b2b1df586fc01e80ecf001770d941385602682ec2055ba7b3979a02c8ebf android.hardware.neuralnetworks@1.3::types
274fb1254a6d1a97824ec5c880eeefc0e410dc6d3a2a4c34052201169d2b7de0 android.hardware.radio@1.5::types
c8e81d912827a5d49b2ddcdc4eb4556c5d231a899a1dca879309e04210daa4a0 android.hardware.radio@1.5::IRadio
a62a93faf173b14a6175b683ebf61ffa568dc61f81e369d2dce7b1265e86cf2f android.hardware.radio@1.5::IRadioIndication

8
neuralnetworks/1.3/types.hal
View file

@ -149,6 +149,7 @@ enum OperationType : int32_t {
* * {@link OperandType::TENSOR_FLOAT16} (since HAL version 1.2)
* * {@link OperandType::TENSOR_FLOAT32}
* * {@link OperandType::TENSOR_QUANT8_ASYMM}
* * {@link OperandType::TENSOR_QUANT8_ASYMM_SIGNED} (since HAL version 1.3)
*
* Supported tensor rank: 4, with "NHWC" or "NCHW" data layout.
* With the default data layout NHWC, the data is stored in the order of:
@ -210,7 +211,8 @@ enum OperationType : int32_t {
* Outputs:
* * 0: The output 4-D tensor, of shape
* [batches, out_height, out_width, depth].
* For a {@link OperandType::TENSOR_QUANT8_ASYMM} tensor,
* For a {@link OperandType::TENSOR_QUANT8_ASYMM} and
* {@link OperandType::TENSOR_QUANT8_ASYMM_SIGNED} tensor,
* the scale and zeroPoint must be the same as input0.
*/
AVERAGE_POOL_2D = @1.2::OperationType:AVERAGE_POOL_2D,
@ -1259,6 +1261,7 @@ enum OperationType : int32_t {
* * {@link OperandType::TENSOR_FLOAT16} (since HAL version 1.2)
* * {@link OperandType::TENSOR_FLOAT32}
* * {@link OperandType::TENSOR_QUANT8_ASYMM}
* * {@link OperandType::TENSOR_QUANT8_ASYMM_SIGNED} (since HAL version 1.3)
*
* Supported tensor rank: 4, with "NHWC" or "NCHW" data layout.
* With the default data layout NHWC, the data is stored in the order of:
@ -1320,7 +1323,8 @@ enum OperationType : int32_t {
* Outputs:
* * 0: The output 4-D tensor, of shape
* [batches, out_height, out_width, depth].
* For a {@link OperandType::TENSOR_QUANT8_ASYMM} tensor,
* For a {@link OperandType::TENSOR_QUANT8_ASYMM} and
* {@link OperandType::TENSOR_QUANT8_ASYMM_SIGNED} tensor,
* the scale and zeroPoint must be the same as input0.
*/
MAX_POOL_2D = @1.2::OperationType:MAX_POOL_2D,