I’ve used PyTorch extensively in many different projects, but had never spent the time to look into uncommon behaviours. A few weeks ago, I encountered a strange bug while using Torch distributed learning, and when I narrowed down the issue, it turned out to be a critical bug. I’m happy to be part of the Torch community now.
If you’ve ever worked with force-directed graphs, you’ve probably heard about Python FA2. However, this package is no longer maintained, and due to changes in Cython for Python 3.9+, it cannot run on newer versions of Python. The good news is that I have made it compatible with Python 3.9+ and published it as a Python package! You can find the source code and instructions to install it here 🙂