Tensorflow example not working on nRF Connect SDK

RE: Tensorflow example not working on nRF Connect SDK

Ivan Angelov — Tue, 13 Jun 2023 09:35:42 GMT

That did the trick!
<my_project>/build/zephyr$ less .config

CONFIG_TENSORFLOW_LITE_MICRO=y

Thank you Sigurd Hellesvik

RE: Tensorflow example not working on nRF Connect SDK

Sigurd Hellesvik — Tue, 13 Jun 2023 09:22:48 GMT

Huh, odd that it is dependent on 0.

You can always try to force set TENSORFLOW_LITE_MICRO in your projects "Kconfig" file:

config TENSORFLOW_LITE_MICRO
    bool "Tensorflow Lite Micro custom override"
    default 1

Regards,
Sigurd Hellesvik

RE: Tensorflow example not working on nRF Connect SDK

Ivan Angelov — Tue, 13 Jun 2023 09:15:46 GMT

Please find attached a screen shot from VSCode studio with nRF Connect plugin:

The option for TensorFlow Lite Micro Support seems to be grayed out.

When I manually set in prj.conf I get following:
warning: TENSORFLOW_LITE_MICRO (defined at modules\tflite-micro\Kconfig:8) was assigned the value
'y' but got the value 'n'. Check these unsatisfied dependencies: 0 (=n). See
docs.zephyrproject.org/.../kconfig.html and/or look up
TENSORFLOW_LITE_MICRO in the menuconfig/guiconfig interface. The Application Development Primer,
Setting Configuration Values, and Kconfig - Tips and Best Practices sections of the manual might be
helpful too.

RE: Tensorflow example not working on nRF Connect SDK

Sigurd Hellesvik — Tue, 13 Jun 2023 08:04:16 GMT

Since the nRF Connect SDK does not have official support available for running tensorflow directly, the help I can give will be limited.
But I will help you look into the issue still, so:

First, can you specify why you

[quote user="Ivan Angelov"]

could not set

CONFIG_TENSORFLOW_LITE_MICRO=y

[/quote]

Try to use our menuconfig tool to search for CONFIG_TENSORFLOW_LITE_MICRO. This will show you which dependencies prevent you to set this config.

Regards,
Sigurd Hellesvik

RE: Tensorflow example not working on nRF Connect SDK

Ivan Angelov — Fri, 09 Jun 2023 10:50:12 GMT

I do not think Nordic have fixed this issue.
I asked for help at tflite-micro repo, see https://github.com/zephyrproject-rtos/zephyr/discussions/52302

I have my own Zephyr project with my own board based on nRF5340. I am using Zephyr v3.2.99-ncs2, nrf-sdk is v2.3.0.
I modified NRF west.yml as suggested in tflite-micro repo:

diff --git a/west.yml b/west.yml
index a93676a6..5b16929e 100644
--- a/west.yml
+++ b/west.yml
@@ -66,6 +66,7 @@ manifest:
         #
         # Please keep this list sorted alphabetically.
         name-allowlist:
+          - tflite-micro
           - TraceRecorderSource
           - canopennode
           - chre

I ran west update but I get no luck with adding CONFIG_TENSORFLOW_LITE_MICRO=y to my project.
When I ran west -v update I noticed following line:

west.manifest: project tflite-micro in file None ignored: an importing manifest blocked or did not allow it

In my project west.yml I have added tflite-micro project:

    - name: tflite-micro
      path: zephyr/modules/tflite-micro
      url: github.com/.../tflite-micro.git
      revision: 8913f033772ac7009b6d8a1c8817e2399d22d803

After west update I can see the sources of tflite-micro located in zephyr repo:

<>/zephyr/modules/tflite-micro
./
../
.bazelrc*
.clang-format*
.editorconfig*
.git/
.github/
.gitignore*
AUTHORS*
BUILD*
CMakeLists.txt*
CODEOWNERS*
CONTRIBUTING.md*
Kconfig*
LICENSE*
README.md*
SECURITY.md*
WORKSPACE*
ci/
data/
docs/
python/
tensorflow/
third_party/

However, I still could not set

CONFIG_TENSORFLOW_LITE_MICRO=y

Are you able to give some tips?

RE: Tensorflow example not working on nRF Connect SDK

kclauah — Wed, 08 Dec 2021 09:52:22 GMT

Sorry. I am not familiar with Edge Impulse.

I misunderstand because the Edge Impulse example KConfig does not include tensorflow lite library. And it can be successfully build and run. So I assume they have their own implementation of ML framework.

Even EON compiler is more optimized than tensorflow lite, I do not think it is reasonable to disable tensorflow. At least, developer has an option to use it or not, right?

Thank you for your patience.

RE: Tensorflow example not working on nRF Connect SDK

Simon — Wed, 08 Dec 2021 07:54:27 GMT

I talked to a developer about this. He informed me that the Edge Impulse (EI) actually uses TensorFlow and it's even stated on their main page (https://www.edgeimpulse.com/)

The EON compiler ued by Edge Impulse is even better than TensorFlow Lite (it's more optimized for embedded systems).

Also, you could use TensorFlow's Keras APIs to edit the ML model in EI studio.

If you want to learn more, take a look at this: https://blog.tensorflow.org/2021/06/how-tensorflow-helps-edge-impulse-make-ml-accessible.html

RE: Tensorflow example not working on nRF Connect SDK

kclauah — Wed, 08 Dec 2021 05:55:47 GMT

So what is the reason of not including tensorflow in the name-allowlist? Edge Impulse is also a machine learning framework but why it does not need the tensorflow lite framework?

RE: Tensorflow example not working on nRF Connect SDK

Simon — Tue, 07 Dec 2021 10:02:54 GMT

Yes, of course, that would be the solution, adding TensorFlow to name-allowlist. I'm sorry for not informing you about this earlier. I tried to add TensorFlow to https://github.com/nrfconnect/sdk-nrf/blob/v1.8.0-rc1/west.yml#L18 and https://github.com/nrfconnect/sdk-nrf/blob/v1.8.0-rc1/west.yml#L96, which I wasn't able to get to work for some reason.

I do not think this is a bug, but rather an active decision by our developers to not include it in the name-allowlist. We do have better support for Edge Impulse samples in the nRF Connect SDK. Check out our Edge Impulse samples here: https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/nrf/samples/edge_impulse/data_forwarder/README.html

Best regards,

Simon

RE: Tensorflow example not working on nRF Connect SDK

kclauah — Tue, 07 Dec 2021 05:14:13 GMT

The tensorflow example works after following the suggestions from @laurenmurphyx64. It is a bug from nrf-sdk and still exists in main stream.

See the GitHub issue for the solutions.

github.com/.../40827

@Simon How can I report a bug here?

RE: Tensorflow example not working on nRF Connect SDK

kclauah — Fri, 03 Dec 2021 09:56:31 GMT

Thank you for the response.

I have followed the getting started guide instructions and install the Zephyr upstream using Ubuntu 20.0.4.

The tensorflow hello_world sample can be compiled correctly using nrf52dk_nrf52832 and nrf9160dk_nrf9160ns.

I have also tried downloading the module from https://github.com/zephyrproject-rtos/tensorflow and put it in <root>\modules\lib\tensorflow but still showing the same error.

Any suggestions is appreciated.

RE: Tensorflow example not working on nRF Connect SDK

Simon — Fri, 03 Dec 2021 08:41:56 GMT

Try installing Zephyr upstream instead of the nRF Connect SDK (NCS), which comes with different packages than NCS (Zephyr west.yml vs. NCS west yml).

Check out this for instructions how to install Zephyr upstream: https://docs.zephyrproject.org/2.7.0/getting_started/index.html

If the Tensorflow sample doesn't work with Zephyr upstream, then you should update https://github.com/zephyrproject-rtos/zephyr/issues/40827 and inform about that.

RE: Tensorflow example not working on nRF Connect SDK

kclauah — Thu, 02 Dec 2021 03:17:02 GMT

Thank you for the suggestions.

I have created an issue in zephyr github.

https://github.com/zephyrproject-rtos/zephyr/issues/40827

But the error

CONFIG_TENSORFLOW_LITE_MICRO was assigned the value y, but got the value n. Missing dependencies: 0

does it mean the tensorflow lite module is missing? I have tried searching all_ops_resolver.h but I got nothing. Maybe I can download the whole repo from https://github.com/tensorflow/tensorflow and put it in somewhere, is it correct?

RE: Tensorflow example not working on nRF Connect SDK

Simon — Wed, 01 Dec 2021 15:01:19 GMT

I can see that CONFIG_TENSORFLOW_LITE_MICRO will get the value 'n', even if it's set to 'y'. I'm not sure why this is. Maybe that's what's leading to the issues you see.

However, since this is a Zephyr sample, I think you'll get better help if you create a Zephyr issue or ask in any of these resources: https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/zephyr/introduction/index.html#resources.

Check out this ticket as well: https://devzone.nordicsemi.com/f/nordic-q-a/79892/tensorflow-lite-micro-support-for-machine-learning/330840#330840

Best regards,

Simon