This project provides high-level wasi-nn bindings for Rust and AssemblyScript. The basic idea: write your machine learning application in a high-level language using these bindings, compile it to WebAssembly, and run it in a WebAssembly runtime that supports the wasi-nn proposal, such as Wasmtime and WasmEdge.
NOTE: These bindings are experimental (use at your own risk) and subject to upstream changes in the wasi-nn specification.
- In Rust, download the crate from crates.io by adding
wasi-nn = "0.6.0"as a Cargo dependency; more information in the Rust README. - In AssemblyScript, download the package from npmjs.com by adding
"as-wasi-nn": "^0.3.0"as an NPM dependency; more information in the AssemblyScript README. - When you call Wasmtime, you'll need to pass the flag
--wasi-modules=experimental-wasi-nnto enable the use use of wasi-nn. - For WasmEdge, you should install the wasi-nn plugin first.
This repository includes examples of using these bindings. See the Rust example and
AssemblyScript example to walk through an end-to-end image classification using an AlexNet model.
Currently the example uses OpenVino as the backend. If you are running Ubuntu, you can simply run
the script to install the supported version.github/actions/install-openvino/install.sh. Otherwise
you'll need to visit the Installation Guides and follow the instructions for your OS. The version
of OpenVino currently supported is openvino_2022.1.0.643.
Once you have OpenVino installed, run them with:
./build.sh rustruns the Rust example./build.sh asruns the AssemblyScript example
To run examples in WasmEdge, consult this article: WasmEdge wasi-nn examples.
This project is licensed under the Apache 2.0 license. See LICENSE for more details.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this project by you, as defined in the Apache-2.0 license, shall be licensed as above, without any additional terms or conditions.