stellargraph compatibility issue with TensorFlow-metal

Has anyone faced any issues in installing "stellargraph" package on Mac mini M1, having TensorFlow-metal framework? This is for Graph Neural Network.

both pip and git based installation failed. (it expects usual TensorFlow 2.x framework already installed). refer to https://stellargraph.readthedocs.io/en/stable/README.html#install-stellargraph-using-pypi for the installation steps/approaches I followed (pip, conda and GitHub).

Any help would be highly appreciated.

Anyone? Please help! Thanks.

Hi @karbapi!

So here I think the issue is that a lot of the packages that check for Tensorflow do it against the package name reserved for the baseline TensorFlow called tensorflow, whereas the TensorFlow package on MacOS is tensorflow-macos. We are working on a generic solution that will fix this issue but it will take until TF2.9 timeline for that to take effect.

In the meantime if you are locally building the stellargraph yourself, you should be able to change the following lines* in setup.py to check for tensorflow-macos instead. Or you can ask for the developers of stellargraph to add a clause there that checks if the OS is MacOS and looks for tensorflow-macos instead of tensorflow.

*https://github.com/stellargraph/stellargraph/blob/3c2c8c18ab4c5c16660f350d8e23d7dc39e738de/setup.py#L25-L27

Thank you very much. I will update asap!

stellargraph compatibility issue with TensorFlow-metal
 
 
Q