📖 About GoMLX
8000GoMLX is a fast and easy-to-use set of Machine Learning and generic math libraries and tools. It can be seen as a PyTorch/Jax/TensorFlow for Go.
It can be used to train, fine-tune, modify and combine machine learning models. It provides all the tools to make that work easy: from a complete set of differentiable operators, all the way to UI tools to plot metrics while training in a notebook.
It runs almost everywhere Go runs, using the a pure Go backend (🚀 NEW 🚀). It runs even in the browser with WASM (see demo created with GoMLX).
It also supports the optimized backend engine based on OpenXLA/PJRT uses just-in-time compilation to CPU and GPU (optionally TPUs also). It's the same engine that powers Google's Jax and TensorFlow, and it has the same speed in many cases. This only compiles for Linux/amd-64 for now (XLA limitation)
Tip
🎓 Quick Start:
- See our tutorial
- See Eli Bendersky's blog post "GoMLX: ML in Go without Python"
- A guided example for Kaggle Dogs Vs Cats.
- Installation here.
It was developed to be full-featured ML platform for Go, and to easily experiment with ML ideas -- see Long-Term Goals below.
It strives to be simple to read and reason about, leading the user to a correct and transparent mental model of what is going on (no surprises) -- aligned with Go philosophy. At the cost of more typing (more verbose) at times.
It is also incredibly flexible, and easy to extend and try non-conventional ideas: use it to experiment with new optimizer ideas, complex regularizers, funky multi-tasking, etc.
Documentation is kept up-to-date (if it is not well documented, it is as if the code is not there) and error messages are useful (always with a stack-trace) and try to make it easy to solve issues.
GoMLX is a full-featured ML framework, supporting various well known ML components
from the bottom to the top of the stack. But it is still only a slice of what a major ML library/framework should provide
(like TensorFlow, Jax or PyTorch).
Examples developed using GoMLX:
- Adult/Census model;
- How KANs learn ?;
- Cifar-10 demo;
- MNIST demo (library and command-line only)
- Dogs & Cats classifier demo;
- IMDB Movie Review demo;
- Diffusion model for Oxford Flowers 102 dataset (generates random flowers);
- Flow Matching Study Notebook based on Meta's "Flow Matching Guide and Code".
- GoMLX/Gemma, a GoMLX implementation of Google DeepMind's Gemma v2 model (blog post)
- GNN model for OGBN-MAG (experimental).
- Last, a trivial synthetic linear model, for those curious to see a barebones simple model.
- 🎉Neural Style Transfer 10 years Celebration🎉: see a demo written using GoMLX of the original paper.
- Triplet Losses: various negative sampling strategies as well as various distance metrics.
- AlphaZero AI for the game of Hive: it uses a trivial GNN to evaluate positions on the board. It includes a WASM demo (runs GoMLX in the browser!) It includes a command-line UI to test your skills!
Highlights:
🚀 NEW 🚀:
SimpleGo, a new Go only backend: it is very portable and doesn't require installation of separate C++ libraries. It is slower and has a limited ops coverage: some 60 ops are already implemented. It is actively being developed, if you are missing any op, please open an issue in GoMLX GitHub and let us know.See amazing Hive demo running on WASM: it evaluates game states using GoMLX
SimpleGocompiled to WebAssembly.
- Converting ONNX models to GoMLX with onnx-gomlx: both as an alternative for
onnxruntime(leveraging XLA), but also to further fine-tune models. See also go-huggingface to easily download ONNX model files from HuggingFace. - Docker "gomlx_jupyterlab" with integrated JupyterLab and GoNB (a Go kernel for Jupyter notebooks)
- Just-In-Time (JIT) compilation using OpenXLA for CPUs and GPUs -- hopefully soon TPUs.
- Autograd: automatic differentiation -- only gradients for now, no jacobian.
- Context: automatic variable management for ML models.
- ML layers library with some of the most popular machine learning "layers": FFN layers,
activation functions, layer and batch normalization, convolutions, pooling, dropout, Multi-Head-Attention (for transformer layers), LSTM, KAN (B-Splines, GR-KAN/KAT networks, Discrete-KAN, PiecewiseLinear KAN), PiecewiseLinear (for calibration and normalization), various regularizations, FFT (reverse/differentiable), learnable rational functions (both for activations and GR-KAN/KAT networks) etc. - Training library, with some pretty-printing. Including plots for Jupyter notebook, using GoNB, a Go Kernel.
- Also, various debugging tools: collecting values for particular nodes for plotting, simply logging the value of nodes during training, stack-trace of the code where nodes are created.
- SGD and Adam (AdamW and Adamax) optimizers.
- Various losses and metrics.
- Pre-Trained models to use: InceptionV3 (image model), many more from HuggingFace using onnx-gomlx. See also go-huggingface to easily download ONNX model files from HuggingFace.
- Support static linking of PJRT: slower to build the Go program, but deploying it doesn't require installing a PJRT plugin in the machine you are deploying it.
Use
go build --tags=pjrt_cpu_staticor includeimport _ "github.com/gomlx/gomlx/backends/xla/cpu/static".
- Q&A and discussions
- Issues
- Random brainstorming on projects: just start a Q&A and I'm happy to meet in discord somewhere or VC.
If you want to use only the SimpleGo backend, simply do import _ "github.com/gomlx/gomlx/backends/simplego" and
you are ready -- it will register itself.
If you want the more advanced/faster XLA backend (only available for Linux at the moment), with support for CUDA, follow below.
TLDR;: Two simple options:
(1) Use the Docker;
(2) Use pre-built binaries (C/C++ libraries) for Linux or MacOS (Darwin, outdated 😞, see warning below). See commands below, or more more details see gopjrt installation instructions.
Linux/amd64, run (see source):
curl -sSf https://raw.githubusercontent.com/gomlx/gopjrt/main/cmd/install_linux_amd64.sh | bashIn addition, for Linux+CUDA (NVidia GPU) support, run (see source):
curl -sSf https://raw.githubusercontent.com/gomlx/gopjrt/main/cmd/install_cuda.sh | bashDepending on what data formats you use, you may want to install hdf5-tools programs (sudo apt install hdf5-tools in Linux).
Use the simplego backend for now.
For XLA unfortunately there are no available versions for Macs.
The easiest to start playing with it, it's just pulling the docker image that includes GoMLX + JupyterLab + GoNB (a Go kernel for Jupyter) and Nvidia's CUDA runtime (for optional support of GPU) pre-installed -- it is ~5Gb to download.
From a directory you want to make visible in Jupyter, do:
For GPU support add the flag
--gpus allto thedocker runcommand bellow.
docker pull janpfeifer/gomlx_jupyterlab:latest
docker run -it --rm -p 8888:8888 -v "${PWD}":/home/jupyter/work janpfeifer/gomlx_jupyterlab:latestIt will display a URL starting with 127.0.0.1:8888 in the terminal (it will include a secret token needed) that you can open in your browser.
You can open and interact with the tutorial from there, it is included in the docker under the directory Projects/gomlx/examples/tutorial.
More details on the docker here.
See the tutorial here. It covers a bit of everything.
After that look at the demos in the examples/ directory.
The library itself is well documented (pls open issues if something is missing), and the code is not too hard to read. Godoc available in pkg.go.dev.
Finally, feel free to ask questions: time allowing (when not in work) I'm always happy to help -- I created groups.google.com/g/gomlx-discuss, or use GitHub discussions page.
Inference or serving a model is done currently by using the Go code used to create the model along with the checkpoint with the trained weights and hyperparameters used to train the model. In other words, it uses the same tools used for training.
For a simple example of how to do this and export a model inference as a library, see
.../examples/cifar/classifer,
and its use in the last cells of the Cifar-10 demo.
In the future we plan to also export models to ONNX or StableHLO and one could use tools that serve those.
- Building and training models in Go -- as opposed to Python (or some other language) -- with focus on:
- Being simple to read and reason about, leading the user to a correct and transparent mental model of what is going on. Even if that means being more verbose when writing.
- Clean, separable APIs: individual APIs should be self-contained and decoupled where possible.
- Composability: Any component should be replaceable, so they can be customized and experimented. That means sometimes more coding (there is not one magic train object that does everything), but it makes it clear what is happening, and it's easy to replace parts with a third party versions or something custom.
- Up-to-date documentation: if the documentation is not there or if it's badly written, it's as if the code was not there either.
- Clear and actionable error reporting
- To be a productive research and educational platform to experiment with new ML ideas and learn.
- Support mirrored training on multiple devices and various forms of distributed training (model and/or data parallelism) in particular to support for large language models and similarly large model training.
- To be a robust and reliable platform for production. Some sub-goals:
- Support modern accelerator hardware like TPUs and GPUs.
- Multiple backends beyond XLA, e.g: llamacpp, WebNN (with Wasm), pure Go version, etc.
- Import pre-trained models from Hugging Face Hub and allow fine-tuning -- ONNX versions already working for many models in onnx-gomlx.
- Compile models to binary as in C-libraries and/or WebAssembly, to be linked and consumed (inference) anywhere (any language).
- What are the environment variables are used by GoMLX ?
GOMLX_BACKEND: defines the backend engine to use (if usingbackends.New()). The value is formatted as "<backend_name>[:<backend_config>]", with the config part being optional. Examples:GOMLX_BACKEND=go: Use theSimpleGobackend, the pure Go implementation that is very portable but slow.GOMLX_BACKEND="xla:cpu": Use XLA (the faster backend, only runs on Linux now) for CPUGOMLX_BACKEND="xla:cuda": Use XLA for for Nvidia CUDAGOMLX=BACKEND="xla:/path/to/my/pjrt_plugin.so": Use XLA with an arbitrary PJRT. PJRT is a plugin system for XLA to support different hardwares. One can install PJRTs build for NVIDIA GPUs (there is an installation script for that), there is also one for ROCm (not tested by the author), for TPU (Google Cloud) and reports of PJRTs being built to even new accelerators (e.g: TensTorrent XLA)
PJRT_PLUGIN_LIBRARY_PATH: the underlying XLA backend uses this variable as an extra directory to search for plugin locations. It searches for the systems library paths ($LD_LIBRARY_PATH,/etc/ld.so.conf), the default/usr/local/lib/gomlx/pjrtand$PJRT_PLUGIN_LIBRARY_PATHif set.
- What backends to include when using GoMLX ?
- The recommendation is to use
import _ "github.com/gomlx/gomlx/backends/default"which will importXLAandSimpleGobackends. If you add-tags=noxlato the compiler it won't include XLA. import _ "github.com/gomlx/gomlx/backends/simplego"to include onlySimpleGo. If you are working on a platform not supported by XLA, or you don't want to install its C++ library.import _ "github.com/gomlx/gomlx/backends/xla"to import only XLA.
- The recommendation is to use
The project is looking forward contributions for anyone interested. Many parts are not yet set in stone, so there is plenty of space for improvements and re-designs for those interested and with good experience in Go, Machine Learning and APIs in general. See the TODO file for inspiration.
No governance guidelines have been established yet.
Copyright 2024 Jan Pfeifer
GoMLX is distributed under the terms of the Apache License Version 2.0. Unless it is explicitly stated otherwise, any contribution intentionally submitted for inclusion in this project shall be licensed under Apache License Version 2.0 without any additional terms or conditions.