rok commented on PR #33925: URL: https://github.com/apache/arrow/pull/33925#issuecomment-1423156047
> > if we assume dimension names are universal > > This is a bonkers thing to assume though, like just get out of batch dimension and image and text data and like, the tensors dimensions can mean literally anything. You're not going to be able to standardize for every case. Indeed :D. And that's just on the application entry/exit. If you want to serialize a tensor somewhere mid computation graph you might as well use UUID for dimension naming. > Is there a third category of use cases? Nothing significant comes to mind. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
