Mark Saroufim

Sounds good, I'll repurpose this PR to delete what we no longer need in `pytorch/pytorch` Tomorrow I'm spending some time looking at improving logging in the fx partitioner

@pytorchbot merge -g

Hi @Narsil it may be nice to have a few `good-first-issue` populated, for one can think of 4 people I work with every day that know a lot about ML...

Am happy to add more comprehensive ones but wasn't too sure how much @erincatto is willing to add to the README given that it's pretty short already.

As an alternative, I've been seeing people in open source leverage Rich to create a dashboard inside of a terminal https://www.willmcgugan.com/blog/tech/post/building-rich-terminal-dashboards/ So maybe you can reduce the burden of creating...

The interesting part is jit gets optimized progressively (albeit in an unstable way) and tops at about 10 iterations and not 1 :confounded: - I don't know enough about JIT...

> aot_function(vmap(f)) does (mostly work), but not the other way around. I think this could work, I believe jax recommendation is to jit the vmap and not vmap the jit...

@zou3519 I'm curious if you have any thoughts on how scheduling would work then. If all the models are being combined into a single output then you're going to waiting...

@robmarkcole would you like to make the PR to fix this?

@rohan-varma @lessw2020 @HamidShojanazeri once you tell @hudeven and I that you'd like to merge the PR let us know. This has been open for a while. Feel free to close...