Commit graph

184 commits

Author SHA1 Message Date
Xusong Wang
c3ea0775ba Add INSTANCE_NORMALIZATION to operation enum.
Bug: 120064464
Test: NeuralNetworksTest_static
Change-Id: Ia81c8e4a79e65275e4a85b93e3bc18a02ac2cbb0
Merged-In: Ia81c8e4a79e65275e4a85b93e3bc18a02ac2cbb0
(cherry picked from commit 897c0afbdf)
2019-01-10 18:08:27 +00:00
Przemyslaw Szczepaniak
2fadc84792 mutateOperationOperandTypeSkip for new CONV_2D variant.
Bug: 119255406
Test: Vts NNAPI tests.
Change-Id: I265ae4968f5fe3589d79e5846a7d640a735a012f
Merged-In: I265ae4968f5fe3589d79e5846a7d640a735a012f
(cherry picked from commit f54f12634c)
2019-01-10 08:16:56 +00:00
Michael K. Sanders
41e6732a66 Adds float16 support for RANDOM_MULTINOMIAL.
Bug: 118607845
Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Change-Id: Iac8c0c23a8219702697a7a126e5e2e43ef0461f4
Merged-In: Iac8c0c23a8219702697a7a126e5e2e43ef0461f4
(cherry picked from commit 5b2615be21)
2019-01-10 08:16:38 +00:00
Przemyslaw Szczepaniak
4766f8b1a1 Add TENSOR_QUANT8_SYMM_PER_CHANNEL to operand types
Added ExtraParams union for extra Operand parameters.
It's a more sensible approach than adding more fields
to the Operand struct.

Bug: 119249581
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_0TargetTest
Test: VtsHalNeuralnetworksV1_1TargetTest
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: I59731134cf0ea34cf9e10342686d331da9e9c3b3
Merged-In: I59731134cf0ea34cf9e10342686d331da9e9c3b3
(cherry picked from commit faa59b8a2c)
2019-01-10 08:15:58 +00:00
Slava Shklyaev
edf67ee8e7 Add REDUCE_* operations
Bug: 113564646
Test: mma
Change-Id: I6088e270c0167bcf8fc6c4002ed7f4e939b6d435
Merged-In: I6088e270c0167bcf8fc6c4002ed7f4e939b6d435
(cherry picked from commit 4fdb8b5727)
2019-01-09 09:45:33 +00:00
David Gross
4592ed15cd Add @1.2::IPreparedModel::executeSynchronously() and corresponding VTS tests.
Bug: 119274127

Test: all of the following, with the appropriate android.hardware.neuralnetworks@1.${X}::IDevice/sample-all
    VtsHalNeuralnetworksV1_0TargetTest
    VtsHalNeuralnetworksV1_0TargetTest
    VtsHalNeuralnetworksV1_1CompatV1_0TargetTest
    VtsHalNeuralnetworksV1_1CompatV1_0TargetTest
    VtsHalNeuralnetworksV1_1TargetTest
    VtsHalNeuralnetworksV1_1TargetTest
    VtsHalNeuralnetworksV1_2CompatV1_0TargetTest
    VtsHalNeuralnetworksV1_2CompatV1_0TargetTest
    VtsHalNeuralnetworksV1_2CompatV1_1TargetTest
    VtsHalNeuralnetworksV1_2CompatV1_1TargetTest
    VtsHalNeuralnetworksV1_2TargetTest
    VtsHalNeuralnetworksV1_2TargetTest

Change-Id: Iedfa485b4008d9cec3b81ff4c0ce3ebc0b83c823
(cherry picked from commit 49e41678f5)
2019-01-02 11:13:11 -08:00
Xusong Wang
47c26b7226 Merge "Add INSTANCE_NORMALIZATION to operation enum." 2018-12-28 04:20:19 +00:00
David Gross
782bb2c095 Merge "Add @1.2::IPreparedModel::executeSynchronously() and corresponding VTS tests." 2018-12-22 02:52:22 +00:00
David Gross
49e41678f5 Add @1.2::IPreparedModel::executeSynchronously() and corresponding VTS tests.
Bug: 119274127

Test: all of the following, with the appropriate android.hardware.neuralnetworks@1.${X}::IDevice/sample-all
    VtsHalNeuralnetworksV1_0TargetTest
    VtsHalNeuralnetworksV1_0TargetTest
    VtsHalNeuralnetworksV1_1CompatV1_0TargetTest
    VtsHalNeuralnetworksV1_1CompatV1_0TargetTest
    VtsHalNeuralnetworksV1_1TargetTest
    VtsHalNeuralnetworksV1_1TargetTest
    VtsHalNeuralnetworksV1_2CompatV1_0TargetTest
    VtsHalNeuralnetworksV1_2CompatV1_0TargetTest
    VtsHalNeuralnetworksV1_2CompatV1_1TargetTest
    VtsHalNeuralnetworksV1_2CompatV1_1TargetTest
    VtsHalNeuralnetworksV1_2TargetTest
    VtsHalNeuralnetworksV1_2TargetTest

Change-Id: Iedfa485b4008d9cec3b81ff4c0ce3ebc0b83c823
2018-12-21 11:25:52 -08:00
Przemyslaw Szczepaniak
47b9141609 mutateOperationOperandTypeSkip for new DEPTHWISE_CONV_2D variant.
Bug: 119255406
Test: Vts NNAPI tests.
Change-Id: Iaeb1ef1095dde51e9324f526a2126ed5d0491586
2018-12-21 11:40:10 +00:00
Xusong Wang
5666626b33 Add new OperandType FLOAT16.
Bug: 120225191
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: Iaabebe490303c9c93a34d319c8c594df13fb7899
Merged-In: Iaabebe490303c9c93a34d319c8c594df13fb7899
(cherry picked from commit 7bca34bee7)
2018-12-18 16:20:00 +00:00
Lev Proleev
a26349a847 Add two missing comparison ops to HAL
Bug: 113561733
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: Ied32f217f4e4ece3c082f6485f99cf8dee78e55f
Merged-In: Ied32f217f4e4ece3c082f6485f99cf8dee78e55f
(cherry picked from commit 149f2b018b)
2018-12-18 09:17:15 +00:00
Michael K. Sanders
2a0101203f Fixes VTS test failures.
Moves comments for type range values into their enums to make it harder
to forget to update them.

Adds missing float16 types for ARGMAX/ARGMIN/CAST.

Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: [  PASSED  ] 2519 tests.

Change-Id: Ic7c3df8c8fbff45fe497f304b6b2c7a09e7dc5a6
Merged-In: Ic7c3df8c8fbff45fe497f304b6b2c7a09e7dc5a6
(cherry picked from commit bbdab2feee)
2018-12-18 09:15:51 +00:00
Xusong Wang
897c0afbdf Add INSTANCE_NORMALIZATION to operation enum.
Bug: 120064464
Test: NeuralNetworksTest_static
Change-Id: Ia81c8e4a79e65275e4a85b93e3bc18a02ac2cbb0
2018-12-17 14:17:46 -08:00
Przemyslaw Szczepaniak
0c157d30f2 Merge "mutateOperationOperandTypeSkip for new CONV_2D variant." 2018-12-14 16:12:16 +00:00
Xusong Wang
1a06e77831 Add 1.2 NN HAL: IPreparedModel & callbacks.
Create 1.2 version IPreparedModel, IPreparedModelCallback, and
IExecutionCallback.

Currently the new interfaces are created the same as 1.0 version,
but will have more methods introduced in later CLs.

Bug: 73506513
Test: VtsHalNeuralnetworksV1_xTargetTest with 1.2 sample driver
Change-Id: Icf4d04c22f88e825d87562f1489377fdf6bf585d
Merged-In: Icf4d04c22f88e825d87562f1489377fdf6bf585d
(cherry picked from commit b5cb8f7632)
2018-12-13 13:56:51 -08:00
Xusong Wang
a3aa40af3e Merge "Add 1.2 NN HAL: IPreparedModel & callbacks." 2018-12-13 20:55:48 +00:00
Slava Shklyaev
51e44f7090 Add new operand type TENSOR_BOOL8
Bug: 119096815
Test: mma
Change-Id: I2adc7e7a328d53a9dafba52a1f8dfaba63e234b2
Merged-In: I2adc7e7a328d53a9dafba52a1f8dfaba63e234b2
(cherry picked from commit 4aad16b684)
2018-12-13 09:18:53 +00:00
Przemyslaw Szczepaniak
f54f12634c mutateOperationOperandTypeSkip for new CONV_2D variant.
Bug: 119255406
Test: Vts NNAPI tests.
Change-Id: I265ae4968f5fe3589d79e5846a7d640a735a012f
2018-12-12 12:54:06 +00:00
Xusong Wang
b5cb8f7632 Add 1.2 NN HAL: IPreparedModel & callbacks.
Create 1.2 version IPreparedModel, IPreparedModelCallback, and
IExecutionCallback.

Currently the new interfaces are created the same as 1.0 version,
but will have more methods introduced in later CLs.

Bug: 73506513
Test: VtsHalNeuralnetworksV1_xTargetTest with 1.2 sample driver
Change-Id: Icf4d04c22f88e825d87562f1489377fdf6bf585d
2018-12-11 17:33:15 -08:00
Przemyslaw Szczepaniak
faa59b8a2c Add TENSOR_QUANT8_SYMM_PER_CHANNEL to operand types
Added ExtraParams union for extra Operand parameters.
It's a more sensible approach than adding more fields
to the Operand struct.

Bug: 119249581
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_0TargetTest
Test: VtsHalNeuralnetworksV1_1TargetTest
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: I59731134cf0ea34cf9e10342686d331da9e9c3b3
2018-12-10 13:08:03 +00:00
Slava Shklyaev
4fdb8b5727 Add REDUCE_* operations
Bug: 113564646
Test: mma
Change-Id: I6088e270c0167bcf8fc6c4002ed7f4e939b6d435
2018-12-10 10:31:15 +00:00
Xusong Wang
2256f080d4 Merge "Add new OperandType FLOAT16." 2018-12-08 02:49:25 +00:00
Keun Soo YIM
68ae05dd2d pack VTS cc_test binaries as general-tests
Test: make general-tests
Bug: 120093339
Merged-In: I363450d205868f900e4925ccff1430e2a569f2a4
Change-Id: I363450d205868f900e4925ccff1430e2a569f2a4
2018-12-07 10:49:56 -08:00
TreeHugger Robot
3ef664747e Merge "Adds float16 support for RANDOM_MULTINOMIAL." 2018-12-07 09:51:22 +00:00
Xusong Wang
7bca34bee7 Add new OperandType FLOAT16.
Bug: 120225191
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: Iaabebe490303c9c93a34d319c8c594df13fb7899
2018-12-06 16:32:28 -08:00
Michael K. Sanders
5b2615be21 Adds float16 support for RANDOM_MULTINOMIAL.
Bug: 118607845
Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Change-Id: Iac8c0c23a8219702697a7a126e5e2e43ef0461f4
2018-12-06 12:35:10 +00:00
Lev Proleev
149f2b018b Add two missing comparison ops to HAL
Bug: 113561733
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: Ied32f217f4e4ece3c082f6485f99cf8dee78e55f
2018-12-05 16:24:38 +00:00
Miao Wang
5a080f41a9 Merge changes from topics "replace_asymm", "fp16-op-add"
* changes:
  Replace TENSOR_QUANT16_ASYMM with TENSOR_QUANT16_SYMM
  Fix VTS ValidationTest for 1.2 ops.
  Adds float16 support to generated tests.
  Autogenerates VTS ValidationTest tests.
  Fix VTS ValidationTest for 1.2 ops.
  Separates VTS tests by HAL version.
2018-12-03 18:48:58 +00:00
Michael K. Sanders
bbdab2feee Fixes VTS test failures.
Moves comments for type range values into their enums to make it harder
to forget to update them.

Adds missing float16 types for ARGMAX/ARGMIN/CAST.

Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: [  PASSED  ] 2519 tests.

Change-Id: Ic7c3df8c8fbff45fe497f304b6b2c7a09e7dc5a6
2018-12-03 17:08:45 +00:00
Slava Shklyaev
61984d221b Merge changes from topic "nnapi-tensor-bool"
* changes:
  Add TENSOR_BOOL8 support to generated tests
  Add new operand type TENSOR_BOOL8
2018-12-03 14:44:38 +00:00
Slava Shklyaev
4aad16b684 Add new operand type TENSOR_BOOL8
Bug: 119096815
Test: mma
Change-Id: I2adc7e7a328d53a9dafba52a1f8dfaba63e234b2
2018-11-30 17:54:55 +00:00
Steven Moreland
4fb17ff880 Merge "Update makefiles." am: 2f3c139981 am: 464839dc43
am: a3beee1be9

Change-Id: I5ff7f3a71d3b2ae503f347ffd50b94797dc92a7e
2018-11-30 01:19:22 -08:00
Steven Moreland
6b110994d2 Update makefiles.
Bug: N/A
Test: N/A
Change-Id: Ie4ccc7f7638e9eca3bcd43c36c16f2f0053ac596
2018-11-29 18:05:35 -08:00
Lev Proleev
217c4071cc Replace TENSOR_QUANT16_ASYMM with TENSOR_QUANT16_SYMM
* Update doc string
* Update zero point mutation to check for symmetric quantization

Fix: 118671831
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: Id1999c793c839b892cfe45cbb245611b12db2a72
Merged-In: Id1999c793c839b892cfe45cbb245611b12db2a72
(cherry picked from commit 48c8820bac)
2018-11-29 18:46:36 +00:00
Xusong Wang
1a2492f311 Fix VTS ValidationTest for 1.2 ops.
Fix VTS ValidationTest for 1.2 ops by skipping invalid tests.

Bug: 119425787

Test: VtsHalNeuralnetworksV1_xTargetTest with 1.2 sample driver
Change-Id: I61e33d8ff3d9121d9ceb28e667ba6c00bb9951ac
Merged-In: I61e33d8ff3d9121d9ceb28e667ba6c00bb9951ac
(cherry picked from commit 5b747ae02b)
2018-11-29 18:46:36 +00:00
Michael K. Sanders
a8efddc558 Autogenerates VTS ValidationTest tests.
This removes the use of a separately updated list of models
that has fallen out of sync.

Bug: 119293899
Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.0::IDevice/sample-all

Change-Id: I2d8804d78331b8fceab4c622c871802aa0f0a4b4
Merged-In: I2d8804d78331b8fceab4c622c871802aa0f0a4b4
(cherry picked from commit b5fe58b95a)
2018-11-29 18:46:36 +00:00
Xusong Wang
a6b1dc7d9b Fix VTS ValidationTest for 1.2 ops.
Skip L2_NORMALIZATION,  LOCAL_RESPONSE_NORMALIZATION, and SOFTMAX for
addOperationInputTest.

Test: VtsHalNeuralnetworksV1_xTargetTest with 1.2 sample driver
Change-Id: I16de09e2eef7610e564f1f49ff067af7efca3784
Merged-In: I16de09e2eef7610e564f1f49ff067af7efca3784
(cherry picked from commit 6433728e2c)
2018-11-29 18:46:36 +00:00
Michael K. Sanders
fc2b10ec86 Separates VTS tests by HAL version.
This prevents a fatal relocation error trying to link
VtsHalNeuralnetworksV1_2TargetTest if it exceeds 1GB.

Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.0::IDevice/sample-all
Bug: 119135172

Change-Id: Id5cfdeddc2f0d810b4eb47084640ae1dbf297ea1
Merged-In: Id5cfdeddc2f0d810b4eb47084640ae1dbf297ea1
(cherry-pick from 0fb3892c93)
2018-11-29 18:46:31 +00:00
Keun Soo YIM
ff84c37bc1 pack VTS cc_test binaries as general-tests
Test: make general-tests
Bug: 120093339
Change-Id: I363450d205868f900e4925ccff1430e2a569f2a4
2018-11-27 16:11:41 -08:00
TreeHugger Robot
c3c3ebad33 Merge "Replace TENSOR_QUANT16_ASYMM with TENSOR_QUANT16_SYMM" 2018-11-16 00:42:40 +00:00
Xusong Wang
9319968f99 Merge "Add two operations to 1.2 nn hal." am: af32d47028 am: 0468984419
am: 1bdc688304

Change-Id: Id53736a0e002f05bdf654cf6478aaab40087a791
2018-11-15 06:42:43 -08:00
Lev Proleev
48c8820bac Replace TENSOR_QUANT16_ASYMM with TENSOR_QUANT16_SYMM
* Update doc string
* Update zero point mutation to check for symmetric quantization

Fix: 118671831
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: Id1999c793c839b892cfe45cbb245611b12db2a72
2018-11-15 10:22:50 +00:00
Michael K. Sanders
9233dbedcc Adds symbolic min/max values for type enums.
This abstracts the boundary values for OperandType and
OperationType to avoid the need to update them in the
model validation functions.

Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all

Change-Id: I39155148d67215e32b4eb1991b885f65d5eeaca8
Merged-In: I39155148d67215e32b4eb1991b885f65d5eeaca8
(cherry-pick from c785d46eb6)
2018-11-15 09:15:06 +00:00
Xusong Wang
5b747ae02b Fix VTS ValidationTest for 1.2 ops.
Fix VTS ValidationTest for 1.2 ops by skipping invalid tests.

Bug: 119425787

Test: VtsHalNeuralnetworksV1_xTargetTest with 1.2 sample driver
Change-Id: I61e33d8ff3d9121d9ceb28e667ba6c00bb9951ac
2018-11-13 15:07:57 -08:00
Michael K. Sanders
32072766b8 Merge "Adds symbolic min/max values for type enums." 2018-11-13 10:02:31 +00:00
Michael K. Sanders
c785d46eb6 Adds symbolic min/max values for type enums.
This abstracts the boundary values for OperandType and
OperationType to avoid the need to update them in the
model validation functions.

Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Change-Id: I39155148d67215e32b4eb1991b885f65d5eeaca8
2018-11-12 11:43:31 +00:00
Xusong Wang
35faff38a7 Add two operations to 1.2 nn hal.
Add ABS and ROI_POOLING to 1.2 nn types.hal

Test: mma
Change-Id: I367e5ea58bcc6bf373fd3a477fbf39c9892a2130
2018-11-09 21:00:56 -08:00
Michael K. Sanders
b5fe58b95a Autogenerates VTS ValidationTest tests.
This removes the use of a separately updated list of models
that has fallen out of sync.

Bug: 119293899
Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.0::IDevice/sample-all

Change-Id: I2d8804d78331b8fceab4c622c871802aa0f0a4b4
2018-11-09 12:30:55 +00:00
Michael K. Sanders
0fb3892c93 Separates VTS tests by HAL version.
This prevents a fatal relocation error trying to link
VtsHalNeuralnetworksV1_2TargetTest if it exceeds 1GB.

Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_2CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_1CompatV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.1::IDevice/sample-all
Test: VtsHalNeuralnetworksV1_0TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.0::IDevice/sample-all
Bug: 119135172
Change-Id: Id5cfdeddc2f0d810b4eb47084640ae1dbf297ea1
2018-11-09 09:57:34 +00:00
Michael K. Sanders
da3bdbc918 Implements RANDOM_MULTINOMIAL CTS/VTS tests.
Bug: 113562775
Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Change-Id: I3d39d056bf5034e01ce9c5c5e3f416a63d7ec8eb
Merged-In: I3d39d056bf5034e01ce9c5c5e3f416a63d7ec8eb
(cherry picked from commit 941d61a62d)
2018-11-05 18:04:09 -08:00
Michael K. Sanders
2b5f5eab59 Merge "Implements RANDOM_MULTINOMIAL CTS/VTS tests." 2018-10-24 15:03:49 +00:00
Michael K. Sanders
941d61a62d Implements RANDOM_MULTINOMIAL CTS/VTS tests.
Bug: 113562775
Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Change-Id: I3d39d056bf5034e01ce9c5c5e3f416a63d7ec8eb
2018-10-23 14:38:49 +01:00
Xusong Wang
6433728e2c Fix VTS ValidationTest for 1.2 ops.
Skip L2_NORMALIZATION,  LOCAL_RESPONSE_NORMALIZATION, and SOFTMAX for
addOperationInputTest.

Test: VtsHalNeuralnetworksV1_xTargetTest with 1.2 sample driver
Change-Id: I16de09e2eef7610e564f1f49ff067af7efca3784
2018-10-22 13:53:31 -07:00
Michael K. Sanders
5dd841253d Adds TENSOR_FLOAT16 operand type.
Bug: 113563458
Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Change-Id: If12ceff428e1b1a90ef99b7353f0df60d4ef8010
Merged-In: If12ceff428e1b1a90ef99b7353f0df60d4ef8010
(cherry picked from commit 19d63453d4)
2018-10-19 09:29:09 +01:00
Lev Proleev
68c8c17467 Add TENSOR_QUANT16_ASYMM to operand types
Add new OperandType::TENSOR_QUANT16_ASYMM.
Add VTS validation for the new type.

Bug: 113561892
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_0TargetTest
Test: VtsHalNeuralnetworksV1_1TargetTest
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: I4f9ed6a33d5d3ec227e9f335df71954c73edf344
Merged-In: I4f9ed6a33d5d3ec227e9f335df71954c73edf344
(cherry picked from commit 5d7c99527e)
2018-10-19 09:29:09 +01:00
Lev Proleev
08662c6f66 Add new OperandType BOOL.
- Add new enum OperandType::BOOL.
- Add v1.2 Operand, OperandType.
- Add VTS validation tests for BOOL.

Bug: 117423393

Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_2TargetTest

Change-Id: I420e2afeb09b881a499eee6b138c1f26e9874f5a
Merged-In: I420e2afeb09b881a499eee6b138c1f26e9874f5a
(cherry picked from commit abad9eac44)
2018-10-19 09:29:09 +01:00
Miao Wang
618028b888 Add HAL entry to allow querying device impl version.
Bug: 111425781
Bug: 112661050
Test: mm
Test: NeuralNetworksTest_static
Change-Id: I32527fa09e45459bc9759f5b679646073cf96785
Merged-In: I32527fa09e45459bc9759f5b679646073cf96785
(cherry picked from commit 44b029b181)
2018-10-19 09:29:09 +01:00
Steven Moreland
2d01a56f97 Merge "Update makefiles." am: c202d17fa6 am: 70d20a497e
am: 4e9e47f47e

Change-Id: I1cce7dda484ddf85556db4c03518463edb29604f
2018-10-17 14:18:52 -07:00
Steven Moreland
1c15915098 Update makefiles.
Some of these were out of date.

Bug: N/A
Test: N/A
Change-Id: I25a2706b95e199f5728cda5a5e4e2b23c60a8d40
2018-10-16 16:36:45 -07:00
Michael K. Sanders
19d63453d4 Adds TENSOR_FLOAT16 operand type.
Bug: 113563458
Test: VtsHalNeuralnetworksV1_2TargetTest --hal_service_instance=android.hardware.neuralnetworks@1.2::IDevice/sample-all
Change-Id: If12ceff428e1b1a90ef99b7353f0df60d4ef8010
2018-10-16 12:59:42 +01:00
Lev Proleev
5d7c99527e Add TENSOR_QUANT16_ASYMM to operand types
Add new OperandType::TENSOR_QUANT16_ASYMM.
Add VTS validation for the new type.

Bug: 113561892
Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_0TargetTest
Test: VtsHalNeuralnetworksV1_1TargetTest
Test: VtsHalNeuralnetworksV1_2TargetTest
Change-Id: I4f9ed6a33d5d3ec227e9f335df71954c73edf344
2018-10-16 10:15:44 +01:00
Xusong Wang
ad43421811 Merge "Add new OperandType BOOL." 2018-10-11 00:45:09 +00:00
Xusong Wang
d4b00fc332 Divide BBOX_TRANSFORM into 2 ops.
Divide BBOX_TRANSFORM op into 2 ops
- AXIS_ALIGNED_BBOX_TRANSFORM
- ROTATED_BBOX_TRANSFORM

Rotated bounding boxes use different tensor format than axis-aligned
bounding boxes, and it would be less confusing if they were represented
by a separate operator code.

Bug: 113562630

Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_xTargetTest with sample driver
Change-Id: Ie08f2e0d0da77f6750766a394969653478d054d5
Merged-In: Ie08f2e0d0da77f6750766a394969653478d054d5
(cherry picked from commit d2bae1c268)
2018-10-10 17:00:48 +00:00
Miao Wang
7b52ab5bf1 Merge "Add HAL entry to allow querying device impl version." 2018-10-09 16:50:33 +00:00
Miao Wang
a57eb5bcd8 Update OWNERS
Test: none
Change-Id: I96db4de3f34837c9f3174fd01e7d06d3355617f3
Merged-In: I96db4de3f34837c9f3174fd01e7d06d3355617f3
(cherry picked from commit 6653e69dc6)
2018-10-09 13:10:11 +00:00
Slava Shklyaev
1079815872 Fix indexing bug in NNAPI
Bug: 114365802
Test: n/a
Change-Id: I87173528b58f66175c99e7c7e2e66daa307b04da
Merged-In: I87173528b58f66175c99e7c7e2e66daa307b04da
(cherry picked from commit 4a211d8364)
2018-10-09 13:10:02 +00:00
Slava Shklyaev
26204621d9 Attach a bug to syncing docs
Bug: 116445845
Test: none
Change-Id: I841b1dc503d70efddb8175b75d9d3efe8078eeb1
Merged-In: I841b1dc503d70efddb8175b75d9d3efe8078eeb1
(cherry picked from commit 7f9f461591)
2018-10-09 13:09:50 +00:00
Przemyslaw Szczepaniak
e9814a9890 Reserve enums for new NNAPI ops.
Test: NeuralNetworksTest_static
Bug: 114365802

Change-Id: I3b379fe3f0336b92ef602bbb12a4bd4a150c573c
Merged-In: I3b379fe3f0336b92ef602bbb12a4bd4a150c573c
(cherry picked from commit 5a9362eb2d)
2018-10-09 13:09:41 +00:00
Slava Shklyaev
48ac2fbc8a Add new op PAD_V2
Bug: 113561890
Test: NeuralNetworksTest_static
Change-Id: I34351d6ff5fae0a13e3e883ca42210c534418ff3
Merged-In: I34351d6ff5fae0a13e3e883ca42210c534418ff3
(cherry picked from commit ffe467e6d8)
2018-10-09 13:09:34 +00:00
Slava Shklyaev
1ba0f52432 Add new ops ARGMAX and ARGMIN
Bug: 113558658
Test: NeuralNetworksTest_static
Change-Id: I8862d4062181689ce69ab65aa07c1b6af07ba567
Merged-In: I8862d4062181689ce69ab65aa07c1b6af07ba567
(cherry picked from commit cc31bd6622)
2018-10-09 13:07:55 +00:00
Lev Proleev
abad9eac44 Add new OperandType BOOL.
- Add new enum OperandType::BOOL.
- Add v1.2 Operand, OperandType.
- Add VTS validation tests for BOOL.

Bug: 117423393

Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_2TargetTest

Change-Id: I420e2afeb09b881a499eee6b138c1f26e9874f5a
2018-10-08 17:21:45 -07:00
Xusong Wang
cf4db1e2d6 Merge "Divide BBOX_TRANSFORM into 2 ops." 2018-10-05 19:47:16 +00:00
Miao Wang
6653e69dc6 Update OWNERS
Test: none
Change-Id: I96db4de3f34837c9f3174fd01e7d06d3355617f3
2018-10-04 17:16:43 -07:00
Xusong Wang
d2bae1c268 Divide BBOX_TRANSFORM into 2 ops.
Divide BBOX_TRANSFORM op into 2 ops
- AXIS_ALIGNED_BBOX_TRANSFORM
- ROTATED_BBOX_TRANSFORM

Rotated bounding boxes use different tensor format than axis-aligned
bounding boxes, and it would be less confusing if they were represented
by a separate operator code.

Bug: 113562630

Test: NeuralNetworksTest_static
Test: VtsHalNeuralnetworksV1_xTargetTest with sample driver
Change-Id: Ie08f2e0d0da77f6750766a394969653478d054d5
2018-10-04 17:15:10 -07:00
Miao Wang
44b029b181 Add HAL entry to allow querying device impl version.
Bug: 111425781
Bug: 112661050
Test: mm
Test: NeuralNetworksTest_static
Change-Id: I32527fa09e45459bc9759f5b679646073cf96785
2018-09-28 10:47:08 -07:00
Slava Shklyaev
4a211d8364 Fix indexing bug in NNAPI
Bug: 114365802
Test: n/a
Change-Id: I87173528b58f66175c99e7c7e2e66daa307b04da
2018-09-27 16:19:38 +01:00
Slava Shklyaev
7f9f461591 Attach a bug to syncing docs
Bug: 116445845
Test: none
Change-Id: I841b1dc503d70efddb8175b75d9d3efe8078eeb1
2018-09-25 16:11:04 +01:00
Przemyslaw Szczepaniak
5a9362eb2d Reserve enums for new NNAPI ops.
Test: NeuralNetworksTest_static
Bug: 114365802

Change-Id: I3b379fe3f0336b92ef602bbb12a4bd4a150c573c
2018-09-25 12:25:13 +00:00
Slava Shklyaev
ffe467e6d8 Add new op PAD_V2
Bug: 113561890
Test: NeuralNetworksTest_static
Change-Id: I34351d6ff5fae0a13e3e883ca42210c534418ff3
2018-09-25 12:24:58 +00:00
Slava Shklyaev
cc31bd6622 Add new ops ARGMAX and ARGMIN
Bug: 113558658
Test: NeuralNetworksTest_static
Change-Id: I8862d4062181689ce69ab65aa07c1b6af07ba567
2018-09-25 10:44:42 +00:00
Slava Shklyaev
af528bc51c Sync docs with NeuralNetworks.h
Bug: 115855152
Test: none
Change-Id: I15b5207f3ffbea57a4c8313ce744a22ce72b402e
2018-09-24 12:31:03 +00:00
Slava Shklyaev
feb87a9cf8 Add VTS tests for NeuralNetworks v1.2
This is a copy the v1.1 tests since we don't have any new ops
implemented in v1.2 yet.

Bug: 114365802
Test: mm
Test: NNAPI VTS
Change-Id: Ida7525fcd3ae0fd6f88ff9591e06aba922bdae64
Merged-In: Ida7525fcd3ae0fd6f88ff9591e06aba922bdae64
(cherry-picked from 871be94770)
2018-09-21 14:46:24 +01:00
Slava Shklyaev
8e139a7a11 Create NeuralNetworks HAL v1.2 for new OperationTypes
Bug: 114365802
Test: mm
Change-Id: I86b9261729a64d02ed30dc822a0226de11473ac8
Merged-In: I86b9261729a64d02ed30dc822a0226de11473ac8
(cherry-picked from 060a9acb3b)
2018-09-21 14:41:55 +01:00