search

google-coral / pycoral

Python API for ML inferencing and transfer-learning on Coral devices
https://coral.ai
Apache License 2.0
347 stars 144 forks source link

Wheels for tflite for python 3.10 #58

Closed rikardn closed 1 year ago

rikardn commented 2 years ago

Description

I don't know if this is the right place to report this, but since Python 3.10 has been released wheels for tflite would be needed.

Click to expand! ### Issue Type Build/Install ### Operating System _No response_ ### Coral Device _No response_ ### Other Devices _No response_ ### Programming Language _No response_ ### Relevant Log Output _No response_
manoj7410 commented 2 years ago

As of now, we don not have a timeline available for Python 3.10 support. However, as soon as we start supporting 3.10, we shall update this issue.

rikardn commented 2 years ago

Thanks for the quick feedback!

johan12345 commented 2 years ago

Looking at the releases page, Python 3.10 is now supported for the Apple M1 platform (macosx_12_0_arm64)... so why not the others as well?

jakew009 commented 2 years ago

Any update on this?

hjonnala commented 2 years ago

Most likely tflite runtime wheels for python 3.10 would be available by end of 2022 by tensorflow/tfltie-support team. https://github.com/tensorflow/tensorflow/issues/56137

rikardn commented 2 years ago

This is good news!

Python 3.11 will be out before this. Do you intend to also publish wheels for Python 3.11 before the year end?

jiapei100 commented 1 year ago

I actually build tflite from source and installed. However, after I successfully built and installed pycoral from source, I got the following ERRORs:

➜  pycoral git:(master) ✗ python3 examples/classify_image.py \                       
--model test_data/mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite \
--labels test_data/inat_bird_labels.txt \
--input test_data/parrot.jpg
Traceback (most recent call last):
  File "....../pycoral/examples/classify_image.py", line 40, in <module>
    from pycoral.utils.edgetpu import make_interpreter
  File "~/.local/lib/python3.10/site-packages/pycoral/utils/edgetpu.py", line 24, in <module>
    from pycoral.pybind._pywrap_coral import GetRuntimeVersion as get_runtime_version
ModuleNotFoundError: No module named 'pycoral.pybind'
hjonnala commented 1 year ago

Please check the Linux machine wheels for tflite and pycoral for python3.10 at: https://github.com/hjonnala/snippets/tree/main/wheels/python3.10

rikardn commented 1 year ago

Thanks! I have tested it and it works fine.

rikardn commented 1 year ago

For information:

This wheel does not work on github actions with ubuntu-latest: ImportError: /lib/x86_64-linux-gnu/libc.so.6: version 'GLIBC_2.33' not found (required by /home/runner/work/pharmpy/pharmpy/.tox/unit/lib/python3.10/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.cpython-310-x86_64-linux-gnu.so)

I also tested ubuntu-22.04 on GHA and it works there.

hjonnala commented 1 year ago

Thanks for the info. It seems ubuntu-latest, on GHA, comes with ubuntu-20.4 and likely to be change to ubuntu-22.04 soon.

hjonnala commented 1 year ago

Please check the Linux machine wheels for tflite and pycoral for python3.10 at: https://github.com/hjonnala/snippets/tree/main/wheels/python3.10

If the wheels are not working for your machine, feel free to try to build for your platform and check the issue to fix the common issues while building: https://github.com/google-coral/pycoral/issues/85#issuecomment-1308139930.. Thanks!

google-coral-bot[bot] commented 1 year ago

Are you satisfied with the resolution of your issue? Yes No

ianustec commented 1 year ago

Please check the Linux machine wheels for tflite and pycoral for python3.10 at: https://github.com/hjonnala/snippets/tree/main/wheels/python3.10

If the wheels are not working for your machine, feel free to try to build for your platform and check the issue to fix the common issues while building: #85 (comment).. Thanks!

Hi, I'm trying to do this (for RPI4) but the process build for python 3.7

` make -C tensorflow/lite/tools/pip_package docker-build \
TENSORFLOW_TARGET=aarch64 PYTHON_VERSION=3.10 docker build -t "tflite-runtime-builder-debian-buster" --build-arg IMAGE=debian:buster . [+] Building 1.5s (15/15) FINISHED
=> [internal] load build definition from Dockerfile 0.0s => => transferring dockerfile: 37B 0.0s => [internal] load .dockerignore 0.0s => => transferring context: 2B 0.0s => [internal] load metadata for docker.io/library/debian:buster 1.3s => [internal] load build context 0.0s => => transferring context: 39B 0.0s => [ 1/10] FROM docker.io/library/debian:buster@sha256:bb8c62f9f540a5f5f49f7ed5caae0e2889f999cd14553f2aaccaf1fa4fcb998d 0.0s => CACHED [ 2/10] COPY update_sources.sh / 0.0s => CACHED [ 3/10] RUN /update_sources.sh 0.0s => CACHED [ 4/10] RUN dpkg --add-architecture armhf 0.0s => CACHED [ 5/10] RUN dpkg --add-architecture arm64 0.0s => CACHED [ 6/10] RUN apt-get update && apt-get install -y debhelper dh-python python-all python-setuptools python-wheel python-numpy python-pip 0.0s => CACHED [ 7/10] RUN pip install pip --upgrade 0.0s => CACHED [ 8/10] RUN pip install pybind11 0.0s => CACHED [ 9/10] RUN pip3 install pip --upgrade 0.0s => CACHED [10/10] RUN pip3 install pybind11 0.0s => exporting to image 0.0s => => exporting layers 0.0s => => writing image sha256:a995ffa6f8be4cb20a68448b69e28290a0a3fd6d75a1fec1362e8a60f856cb5d 0.0s => => naming to docker.io/library/tflite-runtime-builder-debian-buster 0.0s mkdir -p /Users/thomasbenedetti/Downloads/tensorflow/tensorflow/lite/tools/pip_package/out/python3/debian-buster docker run --user 501:20 \ --rm --interactive --tty \ --env "PYTHON=python3" \ --env "TENSORFLOW_TARGET=aarch64" \ --env "BUILD_DEB=n" \ --env "VERSION_SUFFIX=" \ --volume /Users/thomasbenedetti/Downloads/tensorflow/tensorflow/lite/tools/pip_package/../../../..:/tensorflow \ --volume /Users/thomasbenedetti/Downloads/tensorflow/tensorflow/lite/tools/pip_package/out/python3/debian-buster:/out \ "tflite-runtime-builder-debian-buster" \ /bin/bash -c "/tensorflow/tensorflow/lite/tools/pip_package/build_pip_package.sh && \ (cp /tensorflow/tensorflow/lite/tools/pip_package/gen/tflite_pip/.deb \ /tensorflow/tensorflow/lite/tools/pip_package/gen/tflite_pip/python3/dist/{.whl,*.tar.gz} \ /out 2>/dev/null || true)" +++ dirname /tensorflow/tensorflow/lite/tools/pip_package/build_pip_package.sh ++ cd /tensorflow/tensorflow/lite/tools/pip_package ++ pwd

abuzarra commented 1 year ago

Most likely tflite runtime wheels for python 3.10 would be available by end of 2022 by tensorflow/tfltie-support team. tensorflow/tensorflow#56137

Any update about the availability of wheels for Python 3.10?

rikardn commented 1 year ago

At least a wheel for linux on x86 can be found here: https://github.com/hjonnala/snippets/tree/main/wheels/python3.10

abuzarra commented 1 year ago

At least a wheel for linux on x86 can be found here: https://github.com/hjonnala/snippets/tree/main/wheels/python3.10

I saw that wheels, I'm looking for the Linux wheels on aarch64 and armv7l

oberluz commented 10 months ago

At least a wheel for linux on x86 can be found here: https://github.com/hjonnala/snippets/tree/main/wheels/python3.10

I saw that wheels, I'm looking for the Linux wheels on aarch64 and armv7l

Me too. When could we have official wheels for x64, aarch64 and armv7l for python 3.10?

vinzbarbuto commented 5 months ago

Hello everyone, I'm reaching out for an update regarding the official wheels. Currently, I urgently require Python 3.10 to integrate with an external framework. However, I'm encountering a roadblock with pycoral on my Dev Board, which operates on Python 3.7. To address this, I installed Python 3.10, but unfortunately, I'm facing difficulties in downloading the pycoral API as it seems to be unsupported by Python 3.10 at the moment.

I attempted to install pycoral using the provided link for the wheels. Despite successful installation, I encountered the following error message: ImportError: /lib/aarch64-linux-gnu/libc.so.6: version GLIBC_2.32' not found. Upon investigating further by executing ldd --version on my Coral Board, it indicates ldd (Debian GLIBC 2.28-10) 2.28.

Does anyone have any insights or suggestions on how to overcome these challenges?