Jetson nano can't do pip install tensorflow-addons?

How can i do this problem?
Every time see this words, so i’m serching this words , but i can’t find solution.
Could help me please?

sudo pip install tensorflow-addons

ERROR: Could not find a version that satisfies the requirement tensorflow-addons (from versions: none)
ERROR: No matching distribution found for tensorflow-addons

This means that there is no pre-built package for Jetson (arm/aarch64 architecture).

You need to install it from source. Good luck.

Um… one more porblem.
bazel doesn’t work it

bazel build --enable_runfiles build_pip_pkg
bazel-bin/build_pip_pkg artifacts

command not found or no dirctory and name.

bazel 3.4.1 version is not good?
what should i do bazel version?

Each TensorFlow version requires a specific Bazel-Build version, see the build instructions for your TF-version. Btw: building TF on the Nano will take 10+ hours…

The bazel problem was solved.
There was a problem in progress.
Install with cpu option.
The " tensorflow_addons-*.whl" file does not exist in artifacts.
should i just give up?
Its really hard :-(
Thank you for your help.

Output from Ubuntu
SyntaxError: invalid syntax

  • true
  • cp ‘dist/.whl’ /home/bb110/addons/artifacts
    cp: cannot stat 'dist/
    .whl’: No such file or directory

Content inside the “Build_pip_pkg” shell

python ${BUILD_CMD} || true

else

python ${BUILD_CMD} ${NIGHTLY_FLAG} || true

fi

cp dist/*.whl “${DEST}”

popd

rm -rf ${TMPDIR}

echo $(date) : “=== Output wheel file is in: ${DEST}”
}

Hi,

It looks like there is a similar topic on the TensorFlow addons GitHub:
https://github.com/tensorflow/addons/issues/1982

Please check it for information first.
Thanks.

Addons is really simple and speedy to compile

The setup.py just needs a hack for aarm64 but Raspberry Pi 64 OS + Jetson Nano configure.py by Qengineering · Pull Request #2420 · tensorflow/addons · GitHub qengineering is trying to get a pull request through.
But you just need to hack setup.py and then symlink to /usr/lib the pythonlib which setup.py will show location.

If thats a bit confusing use GitHub - Qengineering/TensorFlow-Addons-Jetson-Nano: TensorFlow Addons installation wheels for Jetson Nano