News for hackers. Stuff that matters.
Re:PyTorch 2.0 by chem83 on Friday December 02, 2022 @10:23PM (#33840225)
And ML compilers in the form of IREE/MLIR, which if successful in delivering hw and framework optionality, should make ML frameworks a far less consequential decision.
Re:PyTorch 2.0 by rawrawrawrr on Saturday December 03, 2022 @02:05AM (#33841382)
Not really, because people will still use the dominant framework because that's where all the research, examples, libraries are.