Age | Commit message (Collapse) | Author |
|
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I12f2993d6e7476f9d9cb4590de9ac40bc6dd6449
|
|
* Added Arm NN Build Tool README.md and license.
* Added Dockerfile diagram image and license.
* Updated existing guides: added note to the cross
compile guide to use build-tool, removed content
from the Delegate BuildGuideNative.md.
* Added Arm NN logo image to homepage README.md
and added license.
Signed-off-by: James Conroy <james.conroy@arm.com>
Change-Id: I9306ad24165cf3b0203c2a1852ec3afd718e9eab
|
|
* Added disable-tflite-runtime-fallback option to armnn_delegate
* Updated armnn_delegate version
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I449b16404d3ffe98e6dac52a43e7c25225addd73
|
|
* The previous fix for IVGCVSW-7106 introduced a problem around
operators with multiple inputs and outputs: addSeparator was being
applied to all tensors in the list not just the last one.
Signed-off-by: Colm Donelan <colm.donelan@arm.com>
Change-Id: I0325d9abcb7fb512f834c61686c698bbfc29a3be
|
|
* Move the Conv2D and DepthwiseConv2D validation to Optimization level
when the weights and tensors are as constant inputs
* Take into account offset and scales values when doing INT8 to FP32 dequantization
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I1f81f15640395ac041923b10dbe9151159715117
|
|
Signed-off-by: Cathal Corbett <cathal.corbett@arm.com>
Change-Id: Ib4433767ba8bd3282a9b19c2a85ff1a6e7e83df5
|
|
* Fixed broken links
* Updated link to bazel guide
* Fixed versioning issues in packaging doc
* Updated paths to binary files
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I6120bb2dd56d55a994cd12d99bed054ce4d7c058
|
|
* ProfilingDetails assumed that every workload description included
both tensors and parameters. This is not always the case.
* Modify ProfilingDetails::AddDetailsToString to check the next
element to be printed before deciding to add a separator and new line.
Signed-off-by: Colm Donelan <colm.donelan@arm.com>
Change-Id: I2577b0e8a149d0a172ee12975e18b78238d8256e
|
|
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: Ib91b734d4add47e23ad00f76e53f1873ff617831
|
|
* Fixed the issues about converting Conv2D and DepthwiseConv2d input issues
* Read 1D input tensors that do not have shape specified
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I12f3d1c57a2afedac42c6e7b31e4b1fc689abeca
|
|
* model was declared in the TfLiteExecutor constructor, instead of intializing m_Model
* Working with this model that has 4 output we saw the the output names were not correct, this got fixed too
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I48f194ad4ba6af43d43e6eea336eb87ffee02dcc
|
|
* Adding the check only if it's not a const layer which is needed to run ai_benchmark_v5_yolo_v4_tiny_quant.tflite model
* We still won't be able to run the model due to IVGCVSW-7158
Signed-off-by: Nikhil Raj <nikraj01@e126673.cambridge.arm.com>
Change-Id: Ib7e77a0b5a64be0c92a8e4eae45729f799770b37
|
|
* Adds Dockerfile associated with Arm NN build-tool scripts.
* The Dockerfile encapsulates the installation of system-wide
packages (install-packages.sh), download/install of Arm NN
dependencies (setup-armnn.sh) and the building of Arm NN
and ACL (build-armnn.sh).
* A helper script for copying build contents from the built
Docker image is provided for by docker-copy-to-host.sh.
* Modified existing scripts: moved the cloning of Arm NN and
ACL from setup-armnn.sh to build-armnn.sh and decoupled
setup-armnn.sh from scripts outside of build-tool directory
e.g. armnn/scripts/get_tensorflow.sh.
* The build-armnn.sh script clones the latest release branches
of Arm NN and ACL by default. Custom repos can be placed in
the build-tool directory prior to 'docker build' and they
will be used instead (advanced usage).
* Support added for Linux targets only, Android to be added
in future work.
Signed-off-by: James Conroy <james.conroy@arm.com>
Change-Id: I336013cf93821d2cd3e5d9fe2ca4e955ffdd2386
|
|
* Removed old Dockerfile and associated docs.
* Keeping README as it currently ranks highly on
web searches for "armnn docker" and providing
a link to new alternative Arm NN Build Tool.
Change-Id: I0ce79cedffa035728e394c116912f36481804931
Signed-off-by: James Conroy <james.conroy@arm.com>
|
|
Signed-off-by: Cathal Corbett <cathal.corbett@arm.com>
Change-Id: If00d8dab2846c484a1969fb152cb9f8bd16e1b3e
|
|
Signed-off-by: Nikhil Raj <nikraj01@e126673.cambridge.arm.com>
Change-Id: Ica451b6e4b2f8dd474f1e216d0cb38163a7d2953
|
|
Signed-off-by: Nikhil Raj <nikraj01@e126673.cambridge.arm.com>
Change-Id: I2446013092221f3016f8d2d57256ea0f384d0cec
|
|
* Load the tuning file when GpuAcc is selected as a backend.
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I1d6b0237a87301f07681293a2c6dbc37e352dd28
|
|
Signed-off-by: Nikhil Raj <nikraj01@e126673.cambridge.arm.com>
Change-Id: I849476ac201c151a9c2087bab03af1fc67732bce
|
|
Signed-off-by: Jim Flynn <jim.flynn@arm.com>
Change-Id: I6e5f41a7a4432f448156755389c108692b7213e2
|
|
Signed-off-by: Jim Flynn <jim.flynn@arm.com>
Change-Id: I3a3aab7b5042349cb2df8517678306665e037610
|
|
* dot file to be generated when -v is given. It was only being generated when using the delegate as executor
* output name read from m_Params.m_OutputNames instead of m_TfLiteInterpreter
* typo: "delage" instead of "delegate"
* QAsymmS8 templated as int8, instead of uint8
Change-Id: Ie13ae0f7e6395c0ebcb5ecda32e72082dee8aa6c
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: Iac97a23927ba42290ebeb3446bbd36da15045e07
|
|
Signed-off-by: Nikhil Raj <nikraj01@e126673.cambridge.arm.com>
Change-Id: Ibd6f971f275c115f97a0d8338f6e7770790f2e67
|
|
Signed-off-by: Cathal Corbett <cathal.corbett@arm.com>
Change-Id: I5c68b81a67fc2b5a33cf62753351440564bb868e
|
|
Signed-off-by: Cathal Corbett <cathal.corbett@arm.com>
Change-Id: Ic134f73f217fff420fec089cd817e48ef62641c6
|
|
* Changed long variable declaration to int
Signed-off-by: Samuel Yap <samuel.yap@arm.com>
Change-Id: I2df6f8f6df8780e48e09f7e68c04626a8a8a207d
|
|
* Added case for Bf16 to switch and changed Assertion to Exception
so it shows up in Release build.
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
Change-Id: I817260dc7b7667386c4aa734bea649383866a785
|
|
* Fixed caching issue.
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: Ic7b3e0bd4438b2fd1b3dbfa86b6c89d625bbf9dd
|
|
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: Iecfb4ef0f821f71aaf16c6f3c80f26a4c39f2e3a
|
|
running Arm NN Unittest
Signed-off-by: Cathal Corbett <cathal.corbett@arm.com>
Change-Id: I567452000287babad345e61ea85ea84f362f48e0
|
|
ConvertLayers.
* ConvertBf16ToFp32Layer
* ConvertFp16ToFp32Layer
* ConvertFp32ToBf16Layer
* ConvertFp32ToFp16Layer
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I5e763519a12f017dc14b09ea191fdb3b7398c0d7
|
|
* Originated from a GitHub issue: https://github.com/ARM-software/armnn/issues/667
* Initially, Arm NN supports the pool 2D operation because there is no padding
on the pool2d. Neon failure occurs when padding is followed by average pool 2D
due to folding optimization.
* Here we prevent the folding optimization from happening for the above special case
and add it in as a backend specific optimization.
Signed-off-by: Cathal Corbett <cathal.corbett@arm.com>
Change-Id: Ia0fd90c3a6b4b9d29c81106f154617d2e893e26b
|
|
* Correcting some typos
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: Icb21dc4828e51afa38816bd454926fc41e9e82cb
|
|
* Adding an announcement in the main Readme file with the date we are freezing master
and moving to main branch
Change-Id: I87a0e99d83cc872ef6ac3ba1f0fd445665885059
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
|
|
* Fix compilation error rasied in Nightly_NEW
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I2e7ea907dd2d6e99fc56ff8f48f7c2b6c2a4a7c1
|
|
This reverts commit 1a7f033768acb27da11503bd29abb468d2e77f9e.
List of fixes to be able to add this code again:
* "emplacing_back" the vector inputTensors into the vector m_InputTensorsVec outside the for loop
* GetIOInfo() uses IOptimizedNetwork instead of INetwork, where the infered shapes are not saved
* Add missing data type Signed32 to SetupInputsAndOutputs()
* PrintOutputTensors() prints the actual output without dequantizing
* Add profilingDetailsMethod as input in networkProperties in ArmNNExecutor constructor
* Fix typos
Change-Id: I91de166f87228282db3efa27431fe91458834442
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: Ic6634d48892d11e5f146cdf285e1e333e93e9937
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
|
|
Signed-off-by: Cathal Corbett <cathal.corbett@arm.com>
Change-Id: If24cad1d5d403e195d7adc539afb83cc5df134d1
|
|
* Descriptors added for BatchMatMul
* Layer definition added
* Input validation added (will likely change when opt. param support comes in)
* Ref workload implementation for BatchMatMul added (will also change with opt. param support)
* Ref layer tests made for BatchMatMul
* CMake and other build files updated
Signed-off-by: Samuel Yap <samuel.yap@arm.com>
Change-Id: Ic885301da543ee0fbe7922b85e7f9658c4efc617
|
|
* Initial scripts which download and build Arm NN
dependencies (setup-armnn.sh) and build both
Arm NN and ACL (build-armnn.sh).
* Also added common.sh for handling common variables
between scripts and validation.sh for common
command line arg validation.
Signed-off-by: James Conroy <james.conroy@arm.com>
Change-Id: I6a61149d048f618b4cc9770cc9dd6ec2e5f64ea5
|
|
Fp32NetworkToBf16Converter
* Fuse FP32ToBF16Layers with Constant Layer so Conv2d/FullyConnected
can have their weights redirected.
* If BF16 Unsupported in Conv2d || FullyConnected revert fused
Constant Layer to FP32
Change-Id: If523c708a822659d64597d9ae39cca1c2f84b76f
Signed-off-by: Francis Murtagh <francis.murtagh@arm.com>
|
|
* Fixed model converting issue
* Fixed import memory issue
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: Ied61810b308e0c5d5754f122a6ea2bac1d0725f1
|
|
* Android.bp file gets picked up by older Android versions so
rename it to "switch it off" and update guide on how to enable it
Signed-off-by: Kevin May <kevin.may@arm.com>
Change-Id: I1fc59a0d0dc2e6ce9cdc19f94b0b34d08d414351
|
|
* "Asynchronous Execution with std::launch:async..."
* "Asynchronous Execution with Arm NN thread pool..."
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I93f6ae92fd5599d1042f0dfced7e90ef85e20463
|
|
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: Ia09aba7d8a358f27db7773d360604a2570fdf657
|
|
Signed-off-by: Colm Donelan <colm.donelan@arm.com>
Change-Id: I01000533c3023271f81e9f7f53504ba185a62778
|
|
* Refactor backend capability checks in LoadedNetwork.
* ImportInputs should check the number of tensors does not exceed the
number of inputs.
* In EnqueueWorkload the check for for the count of input tensors
was ignoring pre-imported inputs.
* Added checks to verify ImportInputs/ImportOutputs worked as expected
in EndToEndTestImpl.
* Improve documentation on ImportInputs/ImportOutputs in IRuntime.hpp.
* Disabled import tests in CL and Neon EndToEndTests that cannot work.
Signed-off-by: Colm Donelan <colm.donelan@arm.com>
Change-Id: Iae4b2644a1c9f01ee72bce1afb211661cc9ae2e3
|
|
* Templated the DumpTensor() function based on tensor type
Signed-off-by: Sadik Armagan <sadik.armagan@arm.com>
Change-Id: I08775e480f89010be61daf0a09a2ab0274e05978
|
|
* Remove ARMNN_TF_LITE_DELEGATE and DARMNN_TF_LITE_DELEGATE
Signed-off-by: Teresa Charlin <teresa.charlinreyes@arm.com>
Change-Id: I3fc08da3fa0b733e6791c42f6bc59494f2bc26a6
|
|
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: Id2cf6cda8f83466f76959ca90b03b56864d3ac8a
|
|
Signed-off-by: Nikhil Raj <nikhil.raj@arm.com>
Change-Id: I8129fcd6b20ad6ec3b43f7d3b2bfcb1de45d28e6
|