For what it's worth I was able to successfully install ZenML on Apple Silicon without Rosetta. The steps were roughly:

1. Install ml-metadata by cherry-picking onto the v1.8.0 tag and then following the instructions to build with Bazel.
2. Install ZenML with poetry after removing the markupsafe==1.1.1 dependency.
   I haven't tried to do much with it yet, but this seems like progress.

That's really cool @Adam Merberg!

Awesome, that’s really cool!

@Adam Merberg were you able to proceed beyond this. There are so many people who are fighting with MLMD as we speak

@hamza You mentioned in another thread that there is a non MLMD release coming soon. Does that mean that it would support Apple Silicon natively?

Yes @thomas.kwok thats one of the primary reasons we do it apart from the dependency bloat

Today @Michael Schuster and I are going to sit down on his Mac M1 and try to get ZenML working using @Adam Merberg’s approach though

Glad to hear :+1:

The only other step I was able to take was installing tensorflow-macos in the same environment following . Then I was able to successfully import zenml and tensorflow without any problem. I wasn't able to set up the integration because that wanted to install tensorflow (the non-macos version that presumably wouldn't work).