mediapipe VS Bazel

If you're looking for the differences in terms of how inference is done, I recommend you take a look at MediaPipe's source code. MediaPipe doesn't use raw code, but uses a "graph" instead (eg. pose_landmark_cpu.pbtxt), which can be visualised using MediaPipe Viz. I also used axinc-ai/ailia-models as the base (preprocessing, inference, postprocessing, etc...) which I further built upon (keypoint refinement, roi from keypoints, filtering / smoothing, etc...)

I thought of doing that but unfortunately medipipe requires a RGB input and performs better with it more on that here

You might want to check out https://github.com/open-mmlab/mmhuman3d and https://github.com/google/mediapipe

This architecture was also used in the link referenced when bringing up alternative implementations:

https://github.com/google/mediapipe/issues/877#issuecomment-...

mediapipe: runs in realtime > 20fps confirmed, training code is NOT provided

Install necessary dependencies: sudo apt-get update sudo apt-get upgrade sudo apt-get install build-essential git libhdf5-dev python-dev python-pip python-numpy python-wheel python-mock python-mockito python-pytest python-six python-h5py sudo pip install --upgrade pip sudo pip install setuptools Install Bazel: sudo apt-get install pkg-config zip g++ zlib1g-dev unzip python wget https://github.com/bazelbuild/bazel/releases/download/0.26.1/bazel-0.26.1-installer-linux-x86_64.sh chmod +x bazel-0.26.1-installer-linux-x86_64.sh ./bazel-0.26.1-installer-linux-x86_64.sh --user Clone TensorFlow 1.x: git clone https://github.com/tensorflow/tensorflow.git cd tensorflow git checkout r1.15 Configure TensorFlow: ./configure Build TensorFlow: bazel build --config=opt --local_resources=1024,1.0,1.0 --verbose_failures //tensorflow/tools/pip_package:build_pip_package bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg sudo pip install /tmp/tensorflow_pkg/tensorflow-1.15.0-cp27-cp27mu-linux_armv7l.whl Set up swap memory: sudo fallocate -l 4G /swapfile sudo chmod 600 /swapfile sudo mkswap /swapfile sudo swapon /swapfile Test TensorFlow: python -c "import tensorflow as tf; print(tf.reduce_sum(tf.random.normal([1000, 1000])))"

Bazel's own debian packaging configuration seems to be in scripts/packages/debian. Here's the BUILD file.

There are a better choices out there, bazel is a good example of a really advanced compilation tool that tries to do everything right straight out of the box. The problem is the investment is high. For a large piece of software the migration to bazel can take years. And that doesn't mean all your woes end. Instead you have problems managing the automated parallelism because the problem is still hard: https://github.com/bazelbuild/bazel/issues/10443.

use bazel and hermetic builds, bazel can act as a package manager for the libs you need; it’s not super easy but can be very rewarding once you start caching stuff

If you're interested in going a little further into this, have a look at https://bazel.build to understand how that can all hang together.

- If you can't use 1 or 2, I recommend using bazel (https://bazel.build/) family build system. It's more or less what Google uses internally, and it has great support for multilanguage development.

Oh excellent, then better (and more portable!) tools are available:

http://pants.build

https://ninja-build.org

https://buck.build

and, if you hate yourself: https://bazel.build