Age | Commit message (Collapse) | Author |
|
* Remove permutation when concat axis is inner most
* Add additional parameter to IsMergerSupported as changed in armnn
!armnn:151
Change-Id: Ie214c9573f242d8f04d58fc61621ad3831991d9a
|
|
* Refactored the LSTM unit tests to avoid unnecessary copies
* Added new AddTensorOperand convenience function
* Fixed some required arguments in some tests
* Added new files to Android.bp
!armnn:124
Change-Id: I77ed22f2692d8aa46884b9ee31e48403f7299545
|
|
Change-Id: I15abc32914ddd9c36d9e31ecf40f938304a07fd6
|
|
the ModelToINetworkConverter
* Added Converter function for StridedSlice operator to 1.1:HalPolicy
!armnn:121
Change-Id: Ida911c2ebeae9b7af708a5e4f492e97ac0d336cb
|
|
Change-Id: Id369a2a70863fabaa60ee11849b5449dd95e25a7
|
|
!armnn:117
Depends-On: Ib1c16d871f0898a1caeb6629c1fee6380a773e14
Change-Id: Ie912f018542c69aa9f051e83d3f317145429cca8
|
|
Change-Id: Ie567d55da21ef931ca3ee93fcec61d6196b6b4eb
|
|
* Removes permutes of tensors for Pooling2D,
as NHWC is now supported by Arm NN.
Change-Id: I48417c91f387b6f73bc071e473828f2ee5949332
|
|
the latest CL master (0cf77981fadb883b97efec112e115450873e64c3)
Change-Id: Icbc3a870e058534ff759489b63fbe7bd4502b962
|
|
Change-Id: I9cb4159a4f8e2bccd350bc1a705534203a8edf62
|
|
the HalPolicy
Change-Id: Ibe9ccf814e437c238e474e1f6eac5b1a8db1c8df
|
|
Change-Id: I0e967b46b213a3b122e9d6aa564d909d2c546b00
|
|
!armnn:46
Change-Id: I76177ce9eaa5c1132d3e15158844c2fb4a44d57c
|
|
Change-Id: Iee54c42e9abd927447d958377f3aee266a60ca5d
|
|
Specified NHWC data layout in ResizeBilinear in HAL policy
Change-Id: I0160ec992b595a39ece6d8b060393c5067e8d3a5
|
|
Change-Id: I24fe0916847ec713abf952b9ea911693a692d440
|
|
Mean layer
* Added unit tests (in a new file test/1.1/Mean.cpp)
* Refactored ArmnnDriverImpl to remove redundant tags in the log
messages
* Modified AddInputOperand to accept a quantized input tensor
Change-Id: Ie037ce426777daab28b0501124e1cc8686ad6184
|
|
Change-Id: I12e462c9a0a3613f0158f73929c566f53d1b3c76
|
|
* Refactored HalPolicy::ConvertMean
* Added sanity checks when parsing the Mean operands
Change-Id: Iea19556adb957b200e7fd01022a87230aaaf7157
|
|
* Added the missing CLReduceMean.cpp file to the Android mekfile
Change-Id: Ieaf7c056d6a5e705999a1daef63f5e78d5db3a37
|
|
Change-Id: Ic60c4dacb55bcf2514f011a8e844e7b8f7b13560
|
|
HalPolicy
* Updating SQUEEZE with comments from TRANSPOSE
Change-Id: Iec48c5d583fecc5e72082d7de0fc8b1becfd84d7
|
|
* Removed input/output swizzling from ConvertL2Normalization
Change-Id: I623d878617614aec4d1f41556839808e6184a350
|
|
Change-Id: I5b1209a871d0a848bb1fdfba5228f181e6968004
|
|
Change-Id: I5475932be49419e630627ba042ea6a70151b78e7
|
|
* to auto-register backend factories we need static global varibles
to be initialized which only happens if not omitted by the linker
* the whole-library option makes sure we include the whole ArmnN library
in the driver
!armnn:153399
Change-Id: Idd2fc67ecc848e1b9be75da0e3ceace9bfa0061d
|
|
removing the conversion to splitter layer if pool is 1x1.
Change-Id: Iea3af08652dd6fd1e6820ef43d9101d38c230b5e
|
|
Change-Id: I15dffef32d394b13e57df134000b7dca4b8788af
|
|
* Optimize now returns reasons for failure
!armnn:152481
Change-Id: I1cd7f5796b3c469e9baecbed80e392ac74e40ef4
|
|
!armnn:151777
Change-Id: Ib7eba10f05810c03d7384256828bfd0c23bb6bba
|
|
!armnn:151190
Change-Id: I37eb70543f23b0ef2f6027998413c62784d9b181
|
|
Change-Id: Ie01c1293fc4ecfdf19ef06162a62801c453d3ecd
|
|
* Added L2NormalizationDescriptor to ConvertL2Normalization
!armnn:150051
Change-Id: I622d7a8af2bfd27e7d4543a24c59e53988cb5d82
|
|
Change-Id: I848db966d249a086cd139c407cefced111aa707a
|
|
!builder:149820
Change-Id: I1b3deb81f9dcb976cae6ec4bccc3149e43ae7d28
|
|
Change-Id: I04b6150bb490a25fa7bf2da68c339ca5f1fe75de
|
|
Change-Id: I939cb8d26766c93ee8e3c92909d69549328b0ea7
|
|
Change-Id: I8f90179ed85a8253785b574499c2f9f4898069d4
|
|
* Changed #if defined to #ifdef
* Simplified the Android ML namespace resolution
* Fixed the relative path in some include directives
Change-Id: I46e46faff98559c8042c1a4b8b82007f462df57d
|
|
before and after passing the FP16 flag in the Android-nn-driver
Change-Id: If8d4ca12421c3bee2526eec98f11d393af822373
|
|
* Moved conversion logic into new V1_0 and V1_1 HalPolicy classes
* Extracted common helper functions into ConversionUtils class
Change-Id: I1ab50edc266dd528c0cb22a5cd1aa65e103674d9
|
|
All changes are the same:
//
// Copyright © 2017 ARM Ltd. All rights reserved.
-// See LICENSE file in the project root for full license information.
+// SPDX-License-Identifier: MIT
//
Change-Id: Ibb7ed1e4e2191b9a978e26b5bacf916db403a60e
|
|
Change-Id: I31681c2077f6451dd0a64cae2da7adf92ecad217
|
|
Change-Id: I36f8e7692bcb6664762fc5585c49dcbfa3a46c9d
|
|
refactoring
* Split getCapabilities and getCapabilities_1_1 as it was before
* Setting relaxedFloat32toFloat16Performance when using HAL 1.1 as
required by one of the VTS test
Change-Id: Iff883b8cbd0511596e9848fa40e91e4fa58d4260
|
|
"getSupportedOperations" and "prepareModel" implementations
* Added common base ArmnnDriverImpl class
* Added common template implementation of the driver's "getCapabilities",
"getSupportedOperations" and "prepareModel" methods
* Refactored ArmnnPreparedModel and RequestThread to support HAL
v1.1 models
* Moved "getStatus" to the common base class, as it is shared by both
HAL implementations
* Refactored the code where necessary
Change-Id: I747334730026d63b4002662523fb93608f67c899
|
|
Change-Id: I64fcdb8ac2fd6f9ca0d2811ed2f98008c6396c15
|
|
HAL1.1 operators
Change-Id: I9b10f0a9c88344df108b2325c0233f9660fa6b7c
|
|
ALLOW_MISSING_DEPENDENCIES to true on Android versions that don't
provide all the HAL versions supported by the NN Driver
Change-Id: I7c828649499ca1d7cb26feb93f559e9f1eae33c7
|
|
ModelToINetworkConverter
!armnn:145897
Change-Id: I9684c37bcac148bb5de0ee14d2af331956b511a8
|