How to use the new version of TensorFlow optimized for Apple Silicon M1 on Slicer

Now I have a Macbook Pro with Apple M1, and Operating System is MacOS Big Sur. I found that if I want to use TensorFlow on this Macbook Pro, I need a new version of TensorFlow specifically optimized for the M1 chip, and the download link is GitHub - apple/tensorflow_macos: TensorFlow for macOS 11.0+ accelerated using Apple's ML Compu
However, the Mac-optimized TensorFlow requires Python 3.8 or higher version, and the latest version of Slicer’s Python version is only 3.6.7. Thus the problem of mismatch between them arises. I can’t import tensorflow in Slicer now.
I wonder if there is a solution to this problem. Or has anyone successfully configured TensorFlow for Slicer on an M1 Macbook before?
Looking forward to some guidance! Thank you very much!

That looks cool, but it’s all new. You can try building Slicer with python 3.8 and maybe you can contribute fixes to the the items in the issue linked below.