dm-haiku
equinox
dm-haiku | equinox | |
---|---|---|
10 | 31 | |
2,806 | 1,819 | |
0.9% | - | |
7.8 | 9.2 | |
25 days ago | 14 days ago | |
Python | Python | |
Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
dm-haiku
-
Maxtext: A simple, performant and scalable Jax LLM
Is t5x an encoder/decoder architecture?
Some more general options.
The Flax ecosystem
https://github.com/google/flax?tab=readme-ov-file
or dm-haiku
https://github.com/google-deepmind/dm-haiku
were some of the best developed communities in the Jax AI field
Perhaps the “trax” repo? https://github.com/google/trax
Some HF examples https://github.com/huggingface/transformers/tree/main/exampl...
Sadly it seems much of the work is proprietary these days, but one example could be Grok-1, if you customize the details. https://github.com/xai-org/grok-1/blob/main/run.py
-
Help with installing python packages.
I am fresh to nix os especially when it comes to using python on it how do I install packages withought using pip I need to install numpy~=1.19.5 transformers~=4.8.2 tqdm~=4.45.0 setuptools~=51.3.3 wandb>=0.11.2 einops~=0.3.0 requests~=2.25.1 fabric~=2.6.0 optax==0.0.6 git+https://github.com/deepmind/dm-haiku git+https://github.com/EleutherAI/lm-evaluation-harness/ ray[default]==1.4.1 jax~=0.2.12 Flask~=1.1.2 cloudpickle~=1.3.0 tensorflow-cpu~=2.5.0 google-cloud-storage~=1.36.2 smart_open[gcs] func_timeout ftfy fastapi uvicorn lm_dataformat which I can just do pip -r thetxtfile but idk how to do this in nix os also I would be using python3.7 so far this is what I have come up with but I know its wrong { pkgs ? import {} }: let packages = python-packages: with python-packages; [ mesh-transformer-jax/ jax==0.2.12 numpy~=1.19.5 transformers~=4.8.2 tqdm~=4.45.0 setuptools~=51.3.3 wandb>=0.11.2 einops~=0.3.0 requests~=2.25.1 fabric~=2.6.0 optax==0.0.6 #the other packages ]; pkgs.mkShell { nativeBuildInputs = [ pkgs.buildPackages.python37 ]; }
-
[D] Should We Be Using JAX in 2022?
What's your favorite Deep Learning API for JAX - Flax, Haiku, Elegy, something else?
-
[D] Current State of JAX vs Pytorch?
Just going to add that you should check out haiku if you are considering JAX: https://github.com/deepmind/dm-haiku
-
PyTorch vs. TensorFlow in 2022
As a researcher in RL & ML in a big industry lab, I would say most of my colleagues are moving to JAX 0https://github.com/google/jax], which this article kind of ignores. JAX is XLA-accelerated NumPy, it's cool beyond just machine learning, but only provides low-level linear algebra abstractions. However you can put something like Haiku [https://github.com/deepmind/dm-haiku] or Flax [https://github.com/google/flax] on top of it and get what the cool kids are using :)
-
[D] JAX learning resources?
- https://github.com/deepmind/dm-haiku/tree/main/examples
- Why would I want to develop yet another deep learning framework?
- Help with installing python packages
equinox
-
Ask HN: What side projects landed you a job?
I wrote a JAX-based neural network library (Equinox [1]) and numerical differential equation solving library (Diffrax [2]).
At the time I was just exploring some new research ideas in numerics -- and frankly, procrastinating from writing up my PhD thesis!
But then one of the teams at Google starting using them, so they offered me a job to keep developing them for their needs. Plus I'd get to work in biotech, which was a big interest of mine. This was a clear dream job offer, so I accepted.
Since then both have grown steadily in popularity (~2.6k GitHub stars) and now see pretty widespread use! I've since started writing several other JAX libraries and we now have a bit of an ecosystem going.
[1] https://github.com/patrick-kidger/equinox
-
[P] Optimistix, nonlinear optimisation in JAX+Equinox!
The elevator pitch is Optimistix is really fast, especially to compile. It plays nicely with Optax for first-order gradient-based methods, and takes a lot of design inspiration from Equinox, representing the state of all the solvers as standard JAX PyTrees.
-
JAX – NumPy on the CPU, GPU, and TPU, with great automatic differentiation
If you like PyTorch then you might like Equinox, by the way. (https://github.com/patrick-kidger/equinox ; 1.4k GitHub stars now!)
- Equinox: Elegant easy-to-use neural networks in Jax
- Show HN: Equinox (1.3k stars), a JAX library for neural networks and sciML
-
Pytrees
You're thinking of `jax.closure_convert`. :)
(Although technically that works by tracing and extracting all constants from the jaxpr, rather than introspecting the function's closure cells -- it sounds like your trick is the latter.)
When you discuss dynamic allocation, I'm guessing you're mainly referring to not being able to backprop through `jax.lax.while_loop`. If so, you might find `equinox.internal.while_loop` interesting, which is an unbounded while loop that you can backprop through! The secret sauce is to use a treeverse-style checkpointing scheme.
https://github.com/patrick-kidger/equinox/blob/f95a8ba13fb35...
-
Writing Python like it’s Rust
I'm a big fan of using ABCs to declare interfaces -- so much so that I have an improved abc.ABCMeta that also handles abstract instance variables and abstract class variables: https://github.com/patrick-kidger/equinox/blob/main/equinox/_better_abstract.py
-
[D] JAX vs PyTorch in 2023
For the daily research, I use Equinox (https://github.com/patrick-kidger/equinox) as a DL librarry in JAX.
- [Machinelearning] [D] État actuel de JAX vs Pytorch?
-
Training Deep Networks with Data Parallelism in Jax
It sounds like you're concerned about how downstream libraries tend to wrap JAX transformations to handle their own thing? (E.g. `haiku.grad`.)
If so, then allow me to make my usual advert here for Equinox:
https://github.com/patrick-kidger/equinox
This actually works with JAX's native transformations. (There's no `equinox.vmap` for example.)
On higher-order functions more generally, Equinox offers a way to control these quite carefully, by making ubiquitous use of callables that are also pytrees. E.g. a neural network is both a callable in that it has a forward pass, and a pytree in that it records its parameters in its tree structure.
What are some alternatives?
flax - Flax is a neural network library for JAX that is designed for flexibility.
jax-resnet - Implementations and checkpoints for ResNet, Wide ResNet, ResNeXt, ResNet-D, and ResNeSt in JAX (Flax).
torchtyping - Type annotations and dynamic checking for a tensor's shape, dtype, names, etc.
trax - Trax — Deep Learning with Clear Code and Speed
treex - A Pytree Module system for Deep Learning in JAX
elegy - A High Level API for Deep Learning in JAX
extending-jax - Extending JAX with custom C++ and CUDA code
jax - Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more
diffrax - Numerical differential equation solvers in JAX. Autodifferentiable and GPU-capable. https://docs.kidger.site/diffrax/
jaxline