Jax and Equinox: What are they and why should I bother?

#1

Post by spearman » Mon, Sep 09, 2024, 2:46 PM UTC

Jax and Equinox: What are they and why should I bother?
garymm.org

Re: Jax and Equinox: What are they and why should I bother?

#2

Post by atorodius » Thu, Sep 12, 2024, 6:00 PM UTC

Personally a big fan of Flax. The way it separates the params and the compute graph is - imo - the right (TM) way. Saying this after many years of soing ML :)

Re: Jax and Equinox: What are they and why should I bother?

#3

Post by ubj » Thu, Sep 12, 2024, 9:21 PM UTC

There's quite a few other libraries associated with Equinox in the JAX ecosystem:

https://github.com/patrick-kidger/equinox?tab=readme-ov-file...

I've enjoyed using Equinox and Diffrax for performing ODE simulations. To my knowledge the only other peer library with similar capabilities is the Julia DifferentialEquations.jl package.

Re: Jax and Equinox: What are they and why should I bother?

#4

Post by Lerc » Thu, Sep 12, 2024, 9:26 PM UTC

For me the questions to answer for whether or not I should bother.

Will it try and bind me to other technologies?

Does it work out of the box on ${GPU}?

Is it well supported?

Will it continue to be supported?

Re: Jax and Equinox: What are they and why should I bother?

#5

Post by mccoyb » Thu, Sep 12, 2024, 9:28 PM UTC

Equinox has great idioms — really pioneered the Pytree perspective. Penzai is also great.

JAX feels close to achieving a sort of high-level GPGPU ecosystem. Super fledgling — but I keep finding more little libraries that build on JAX (and can be compositionally used with other JAX libraries because of it).

Only problem is that lots of compositional usage leads to big code and therefore big compile times for XLA.

Re: Jax and Equinox: What are they and why should I bother?

#6

Post by crazygringo » Thu, Sep 12, 2024, 11:59 PM UTC

Lerc wrote: post #4

For me the questions to answer for whether or not I should bother. Will it try and bind me to other technologies? Does it work out of the box on ${GPU}? Is it well supported? Will it continue to be supported?

Playing with JAX on Google Colab (Nvidia T4), everything works great.

Sadly, I cannot get JAX to work with the built-in GPU on my M1 MacBook Air. In theory it's supposed to work:

https://developer.apple.com/metal/jax/

But it crashes Python when I try to run a compiled function. And that's only after discovering I need an older specific version of jax-metal, because newer versions apparently don't work with M1 anymore (only M2/M3) -- they don't even report the GPU as existing. And even if you get it running, it's missing support for complex numbers.

I'm not clear whether it's Google or Apple who is building/maintaining support for Apple M chips though.

JAX works perfectly in CPU mode though on my MBA, so at least I can use it for development and debugging.

Re: Jax and Equinox: What are they and why should I bother?

#7

Post by p1esk » Fri, Sep 13, 2024, 3:39 AM UTC

I wish Jax had everything I need to experiment with DL models built in, natively - like Pytorch. Instead there are many third party libraries (flax, trax, haiku, this one, etc). I have no idea which one to use. This was the case when I first played with jax 5 years ago, and it’s still the case today (even worse it seems). This makes it a non starter for me.

Re: Jax and Equinox: What are they and why should I bother?

#8

Post by spearman » Fri, Sep 13, 2024, 4:26 AM UTC

crazygringo wrote: post #6

Lerc wrote: post #4

For me the questions to answer for whether or not I should bother. Will it try and bind me to other technologies? Does it work out of the box on ${GPU}? Is it well supported? Will it continue to be supported?

Playing with JAX on Google Colab (Nvidia T4), everything works great. Sadly, I cannot get JAX to work with the built-in GPU on my M1 MacBook Air. In theory it's supposed to work: https://developer.apple.com/metal/jax/ But it crashes Python when I try to run a compiled function. And that's only after discovering I need an older specific version of jax-metal, because newer versions apparently don't work with M1 anymore…

Pretty sure it's Apple building it, and they're using JAX in-house so I imagine it will get better over time. Though they do love to drop support for old things so maybe M1 will never work again...

Re: Jax and Equinox: What are they and why should I bother?

#9

Post by spearman » Fri, Sep 13, 2024, 4:30 AM UTC

Lerc wrote: post #4

For me the questions to answer for whether or not I should bother. Will it try and bind me to other technologies? Does it work out of the box on ${GPU}? Is it well supported? Will it continue to be supported?

Re support: JAX is open source and X, Apple, and Google all use it, so I can't imagine it being abandoned in the next 5 years at least.

Re: Jax and Equinox: What are they and why should I bother?

#10

Post by mccoyb » Fri, Sep 13, 2024, 10:28 AM UTC

spearman wrote: post #8

crazygringo wrote: post #6

Earlier quoted context omitted.

Playing with JAX on Google Colab (Nvidia T4), everything works great. Sadly, I cannot get JAX to work with the built-in GPU on my M1 MacBook Air. In theory it's supposed to work: https://developer.apple.com/metal/jax/ But it crashes Python when I try to run a compiled function. And that's only after discovering I need an older specific version of jax-metal, because newer versions apparently don't work with M1 anymore…

Pretty sure it's Apple building it, and they're using JAX in-house so I imagine it will get better over time. Though they do love to drop support for old things so maybe M1 will never work again...

I think they’re likely using MLX in house now, no? (Probably not everyone, ofc - but seems likely that many will just use the native array framework designed explicitly for MX chips)