mingmingtasd

I have validated that this issue will be fixed after integrating nGraph ops.

Submitted [issue#1756](https://github.com/openvinotoolkit/openvino/issues/1756) on openvinotoolkit/openvino.

Tested on google pixel 3.

ONNX ops | polyfill | Android NNAPI | remarks -- | -- | -- | -- Transpose | √ | [ANEURALNETWORKS_TRANSPOSE](https://developer.android.com/ndk/reference/group/neural-networks#group___neural_networks_1ggaabbe492c60331b13038e39d4207940e0a92d7bc95eb68525334b6cfe80cd271ee) | - Pad | √ | [ANEURALNETWORKS_PAD](https://developer.android.com/ndk/reference/group/neural-networks#group___neural_networks_1ggaabbe492c60331b13038e39d4207940e0aaced01fc41e401b81cefcf53780558d1) | -...

Ops tables for **Android NNAPI** and **OpenVINO** have been added as above , please review, thanks! @huningxin

> > Ops tables for **Android NNAPI** > > The `ANEURALNETWORKS_MEAN` op supports `ReduceMean`, does it? Yes, I added it, thanks! @huningxin

Since I tried many times but failed to enable APIs level 29, I submitted an issue to Google Issue Track: https://issuetracker.google.com/issues/146982053.

Under the issue https://issuetracker.google.com/issues/144381394, I asked two questions relevant to creating new executions and using burstCompute to Google and paste the content as below: Mingming1 Xu #8Dec 30, 2019 04:41PM...

I tried to implement this two APIs [ANeuralNetworksExecution_burstCompute](https://developer.android.com/ndk/reference/group/neural-networks#aneuralnetworksexecution_burstcompute) and [ANeuralNetworksExecution_compute](https://developer.android.com/ndk/reference/group/neural-networks.html#aneuralnetworksexecution_compute) and tested the inference times with workload on Pixel 3 and Samsung Note 10 . Compared with our implementation of...

Hi, do you have any ideas? If no, I will close this issue. Thanks! @fujunwei