wonnx
naga
Our great sponsors
wonnx | naga | |
---|---|---|
18 | 28 | |
1,487 | 1,491 | |
6.8% | 0.7% | |
6.5 | 9.2 | |
27 days ago | 6 months ago | |
Rust | Rust | |
GNU General Public License v3.0 or later | GNU General Public License v3.0 or later |
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.
wonnx
-
Intel CEO: 'The entire industry is motivated to eliminate the CUDA market'
The two I know of are IREE and Kompute[1]. I'm not sure how much momentum the latter has, I don't see it referenced much. There's also a growing body of work that uses Vulkan indirectly through WebGPU. This is currently lagging in performance due to lack of subgroups and cooperative matrix mult, but I see that gap closing. There I think wonnx[2] has the most momentum, but I am aware of other efforts.
[1]: https://kompute.cc/
[2]: https://github.com/webonnx/wonnx
-
VkFFT: Vulkan/CUDA/Hip/OpenCL/Level Zero/Metal Fast Fourier Transform Library
To a first approximation, Kompute[1] is that. It doesn't seem to be catching on, I'm seeing more buzz around WebGPU solutions, including wonnx[2] and more hand-rolled approaches, and IREE[3], the latter of which has a Vulkan back-end.
[1]: https://kompute.cc/
[2]: https://github.com/webonnx/wonnx
[3]: https://github.com/openxla/iree
-
Onnx Runtime: “Cross-Platform Accelerated Machine Learning”
There's also a third-party WebGPU implementation: https://github.com/webonnx/wonnx
-
Are there any ML crates that would compile to WASM?
By experimental I meant e.g. using WGPU to run compute shaders like wonnx, which is working fine but only on a very restricted set of devices and browsers.
- WebGPU ONNX inference runtime written in Rust
-
PyTorch Primitives in WebGPU for the Browser
https://news.ycombinator.com/item?id=35696031 ... TIL about wonnx: https://github.com/webonnx/wonnx#in-the-browser-using-webgpu...
microsoft/onnxruntime: https://github.com/microsoft/onnxruntime
Apache/arrow has language-portable Tensors for cpp: https://arrow.apache.org/docs/cpp/api/tensor.html and rust: https://docs.rs/arrow/latest/arrow/tensor/struct.Tensor.html and Python: https://arrow.apache.org/docs/python/api/tables.html#tensors https://arrow.apache.org/docs/python/generated/pyarrow.Tenso...
Fwiw it looks like the llama.cpp Tensor is from ggml, for which there are CUDA and OpenCL implementations (but not yet ROCm, or a WebGPU shim for use with emscripten transpilation to WASM): https://github.com/ggerganov/llama.cpp/blob/master/ggml.h
Are the recommendable ways to cast e.g. arrow Tensors to pytorch/tensorflow?
FWIU, Rust has a better compilation to WASM; and that's probably faster than already-compiled-to-JS/ES TensorFlow + WebGPU.
What's a fair benchmark?
-
rustformers/llm: Run inference for Large Language Models on CPU, with Rust 🦀🚀🦙
wonnx has done some fantastic work in this regard, so that's where we plan to start once we get there. In terms of general discussion of alternate backends, see this issue.
-
I want to talk about WebGPU
> GPU in other ways, such as training ML models and then using them via an inference engine all powered by your local GPU?
Have a look at wonnix https://github.com/webonnx/wonnx
A WebGPU-accelerated ONNX inference run-time written 100% in Rust, ready for native and the web
-
Chrome Ships WebGPU
Looking forward to your WebGPU ML runtime! Also, why not contribute back to WONNX? (https://github.com/webonnx/wonnx)
-
OpenXLA Is Available Now
You can indeed perform inference using WebGPU (see e.g. [1] for GPU-accelerated inference of ONNX models on WebGPU; I am one of the authors).
The point made above is that WebGPU can only be used for GPU's and not really for other types of 'neural accelerators' (like e.g. the ANE on Apple devices).
[1] https://github.com/webonnx/wonnx
naga
- How does webgpu planning to use webgl shaders?
-
I want to talk about WebGPU
That wouldn't have been all that different from WGSL though, the most important thing is that whatever WebGPU uses for its shaders can be translated to and from SPRIV (and WGSL does that too (e.g. via https://dawn.googlesource.com/tint and https://github.com/gfx-rs/naga).
-
Survey: How have shader compilation messages been for you?
Hey all, wanted to put this link in here, where I'm proposing changing the API for errors in naga, so Naga can take ownership of error presentation and actually Make Shader Compilation Messages Comfy™: https://github.com/gfx-rs/naga/issues/2317
-
Start project on Metal, port to DX11?
EDIT: There is also naga but it does not take HLSL as input: https://github.com/gfx-rs/naga but you can use DirectXShaderCompiler to compile to SpirV, then use naga to compile to Metal.
-
Chrome ships WebGPU (available by default in Chrome 113)
And it seems that naga https://github.com/gfx-rs/naga Already has a working front/backend for wgsl.
-
Ray query example in Blade
This is basically Ray Tracing support in Blade. So far, only ray queries are supported. Unlike prior work on ray tracing in Rust, this is original due to all shader code being WGSL, see the Naga PR.
-
Does WGSL work well with vulkan?
There's a compiler that can translate from WGSL to SPIR-V called naga. Having such a compiler is essential, since WebGPU is planned to use WGSL and browsers are expected to implement rendering via Vulkan (and probably Metal and DX12).
-
Glsl transpiler, interpreter?
Not sure about on the CPU, but naga is a shading language transpiler you can write custom front/backends for.
-
Any guides/documentation on the WGSL shading language?
The spec docs are actually pretty useful https://www.w3.org/TR/WGSL/ besides that I was using naga's tests for reference https://github.com/gfx-rs/naga/tree/master/tests
-
How are Vulkan, CUDA, Triton and all other things connected?
For cross-platform support look at WebGPU and Vulkan (e.g,: [0] [1]. Essentially, you would need to write the func in WGSL or GLSL, HLSL or MSL. Each of these can be cross-compiled to SPIR-V (what Vulkan needs) with cross-compilers such as spirv-cross and naga.
What are some alternatives?
stablehlo - Backward compatible ML compute opset inspired by HLO/MHLO
wgsl-cheat-sheet - Cheat sheet for WGSL syntax for developers coming from GLSL.
onnx - Open standard for machine learning interoperability
shaderc - A collection of tools, libraries, and tests for Vulkan shader compilation.
tract - Tiny, no-nonsense, self-contained, Tensorflow and ONNX inference
wgsl.vim - WGSL syntax highlight for vim
iree - A retargetable MLIR-based machine learning compiler and runtime toolkit.
wgsl-mode - Emacs syntax highlighting for the WebGPU Shading Language (WGSL)
burn - Burn is a new comprehensive dynamic Deep Learning Framework built using Rust with extreme flexibility, compute efficiency and portability as its primary goals.
gpuweb - Where the GPU for the Web work happens!
blaze - A Rustified OpenCL Experience
vscode-wgsl - VsCode Syntax highlight for WGSL files