Reading "Use the Object Detection API" causes incompatibility of PyYaml with Cython

[James_J_Johnson]

At the moment, “Reading: Use the Object Detection API” has technical issues. It does not work in Colab and on my laptop (both Windows Power Shell and WSL Ubunty). Here are the links that explain the reason for the failures:

‘Error installing Pyyaml==5.4, Cython_sources · Issue #724 · yaml/pyyaml · GitHub’

‘PyYaml 5.3.1 to 6.0.0 broken with Cython 3.0 release, causing OctoPrint's install to fail · Issue #4854 · OctoPrint/OctoPrint · GitHub’

In summary, the following bash command attempts to install “tf-models-official>=2.5.1”, which requires “pyyaml<6.0,>=5.1”:
python -m pip install .

Sadly, PyYaml versions from 5.4.0 to 6.0.0 are not compatible to its dependency Cython in version 3. Said version 3 was just released on July 17th (yesterday), which now causes installations to fail (see 2nd link above).

The proposed solution is to upgrade PyYaml to 6.0.1, which pins Cython to a version below 3.0.0. But I have no idea how to do this.

PS: I used this link to install relevant packages on Windows, when I used protoc-3.12.3-win64.zip:
https://tensorflow-object-detection-api-tutorial.readthedocs.io/en/latest/install.html#tensorflow-object-detection-api-installation

[James_J_Johnson]

The bad news is that this technical issue is applicable to “Reading: Eager Few Shot Object Detection” as well. In other words, this Colab tutorial does not work for the same reason now (07/18/2023).

[James_J_Johnson]

The good news is that the graded assignment “Zombie Detector” worked for me. It means the solution for the non-working ungraded assignments can be probably found in the above graded assignment.