Wonnx is a GPU-accelerated ONNX inference run-time written 100% in Rust, ready for the web.
API | Windows | Linux & Android | macOS & iOS |
---|---|---|---|
Vulkan | ✅ | ✅ | |
Metal | ✅ | ||
DX12 | ✅ (W10 only) | ||
DX11 | 🚧 | ||
GLES3 | 🆗 |
✅ = First Class Support — 🆗 = Best Effort Support — 🚧 = Unsupported, but support in progress
Ensure your system supports either Vulkan, Metal or DX12 for access to the GPU. Then either download a binary release,
or install Rust and run cargo install --git https://github.com/webonnx/wonnx.git wonnx-cli
to install the CLI.
The CLI tool (nnx
) provides a convenient interface for tinkering with models (see the README for more information):
nnx info ./data/models/opt-squeeze.onnx
nnx infer ./data/models/opt-squeeze.onnx -i data=./data/images/pelican.jpeg --labels ./data/models/squeeze-labels.txt --top 3
Add the wonnx
crate as dependency (cargo add wonnx
if you have cargo-add). Then, see the examples
for usage examples, or browse the API docs.
pip install wonnx
And then, to use:
from wonnx import Session
session = Session.from_path(
"../data/models/single_relu.onnx"
)
inputs = {"x": [-1.0, 2.0]}
assert session.run(inputs) == {"y": [0.0, 2.0]}
Then run python3
with the above Python code!
For more details on the Python package including build instructions, see wonnx-py.
npm install @webonnx/wonnx-wasm
And then, on the client side:
import init, { Session, Input } from "@webonnx/wonnx-wasm";
// Check for WebGPU availability first: if(navigator.gpu) { .. }
await init();
const session = await Session.fromBytes(modelBytes /* Uint8Array containing the ONNX file */);
const input = new Input();
input.insert("x", [13.0, -37.0]);
const result = await session.run(input); // This will be an object where the keys are the names of the model outputs and the values are arrays of numbers.
session.free();
input.free();
The package @webonnx/wonnx-wasm provides an interface to WONNX, which is included as WebAssembly module and will use the browser's WebGPU implementation. See wonnx-wasm-example for a more complete usage example involving a bundler.
For more details on the JS/WASM package including build instructions, see wonnx-wasm.
To work on wonnx itself, follow the following steps:
- Install Rust
- Install Vulkan, Metal, or DX12 for the GPU API.
- git clone this repo.
git clone https://github.com/webonnx/wonnx.git
Then, you're all set! You can run one of the included examples through cargo:
cargo run --example squeeze --release
- To run an onnx model, first simplify it with
nnx prepare
(substitute withcargo run -- prepare
when inside this repo):
nnx prepare -i ./some-model.onnx ./some-model-prepared.onnx
To specify dynamic dimension parameters, add e.g. --set batch_size=1
.
You can also use an external tool, such as onnx-simplifier, with the command:
# pip install -U pip && pip install onnx-simplifier
python -m onnxsim mnist-8.onnx opt-mnist.onnx
- Then you can run it using the CLI (see README or programmatically, following the examples in the examples folder. To run an example:
cargo run --example mnist --release
- Squeezenet
- MNIST
- BERT
Except when running in WebAssembly, you may set the following environment variables to influence GPU selection by WGPU:
WGPU_ADAPTER_NAME
with a substring of the name of the adapter you want to use (e.g.1080
will matchNVIDIA GeForce 1080ti
).WGPU_BACKEND
with a comma separated list of the backends you want to use (vulkan
,metal
,dx12
,dx11
, orgl
).WGPU_POWER_PREFERENCE
with the power preference to choose when a specific adapter name isn't specified (high
orlow
)
Contributions are very much welcomed even without large experience in DL, WGSL, or Rust. I hope that this project can be a sandbox for all of us to learn more about those technologies beyond this project's initial scope.
To implement an operator all you have to do is:
- Add a new matching pattern in
compiler.rs
- Retrieve its attributes values using the
get_attribute
function:
let alpha = get_attribute("alpha", Some(1.0), node);
// or without default value
let alpha = get_attribute::<f32>("alpha", None, node);
- Add any variable you want to use in the WGSL shader using
context
. - Write a new WGSL template in the
templates
folder.
Available types are in
structs.wgsl
but you can also generate new ones within your templates.
- Respect the binding layout that each entry is incremented by 1 starting from 0, with input first and output last. If the number of binding is above 4. Increment the binding group. You can change the input within
sequencer.rs
- Write the logic.
There is default variables in the context:
{{ i_lens[0] }}
: the length of the input 0. This also work for output:{{ o_lens[0] }}
and other input{{ i_lens[1] }}
{{ i_shape[0] }}
: the array of dimensions of input 0. To get the first dimension of the array, just use:{{ i_shape[0][0] }}
{{ i_chunks[0] }}
: the size of the chunks of each dimensions of input 0. By default, each variable is represented as a long array of values where to get to specific values you have to move by chunks. Those chunks are represented within this variable. To get the size of the chunks of the first dimensions use:{{ i_chunks[0][0] }}
.{{ op_type }}
the op type as some op_type like activation are using the same template.
- Test it using the utils function and place it in the tests folder. The test can look as follows:
#[test]
fn test_matmul_square_matrix() {
// USER INPUT
let n = 16;
let mut input_data = HashMap::new();
let data_a = ndarray::Array2::eye(n);
let mut data_b = ndarray::Array2::<f32>::zeros((n, n));
data_b[[0, 0]] = 0.2;
data_b[[0, 1]] = 0.5;
let sum = data_a.dot(&data_b);
input_data.insert("A".to_string(), data_a.as_slice().unwrap());
input_data.insert("B".to_string(), data_b.as_slice().unwrap());
let n = n as i64;
let model = model(graph(
vec![tensor("A", &[n, n]), tensor("B", &[n, n])],
vec![tensor("C", &[n, n])],
vec![],
vec![],
vec![node(vec!["A", "B"], vec!["C"], "MatMul", "MatMul", vec![])],
));
let session =
pollster::block_on(wonnx::Session::from_model(model)).expect("Session did not create");
let result = pollster::block_on(session.run(input_data)).unwrap();
// Note: it is better to use a method that compares floats with a tolerance to account for differences
// between implementations; see `wonnx/tests/common/mod.rs` for an example.
assert_eq!((&result["C"]).try_into().unwrap(),sum.as_slice().unwrap());
}
Check out tera documentation for other templating operation: https://tera.netlify.app/docs/
- If at any point you want to do optimisation of several nodes you can do it within
sequencer.rs
.
Supported Operators (ref ONNX IR)
-
The
Clip
,Resize
,Reshape
,Split
,Pad
andReduceSum
ops accept (typically optional) secondary inputs to set various parameters (i.e. axis). These inputs are only supported if they are supplied as initializer tensors (i.e. do not depend on inputs and are not outputs of other ops), because wonnx pre-compiles all operations to shaders in advance (and must know these parameters up front). -
Internally 64-bit integers are not supported (the reason is they are not supported in the current version of WGSL); inputs and initializers with 64-bit scalars are converted to 32-bit values (possibly overflowing).
-
For
MatMul
andGemm
, the matrix dimensions must be divisible by 2, or the output matrix must be of size (1, N). Matrix multiplication only supports floats, not integers (this is a WebGPU/WGSL limitation).
WONNX needs to know the shape of input and output tensors for each operation in order to generate shader code for executing it. ONNX models however do not always contain this information for intermediate values. Shape inference is the process of deducing the shape of intermediate values from the shape of inputs and outputs and the characteristics of each operation.
WONNX supports a limited form of shape inference (the process of determining what the shapes are of the various nodes in a model's graph). Shape inference is available programmatically as well as through the CLI. Before shape inference can be performed, all dynamic dimension parameters need to be replaced with static values. Shape inference only infers output shapes from input shapes for specific supported ops (see the table above). Inference cannot succeed if the shape for any input of a node is not known. Nodes that already have fully defined shapes for their outputs are left unchanged (and the outputs are used for shape inference on nodes that use these outputs as inputs).
To perform shape inference using the CLI, run a command similar to this (here batch_size
and sequence_length
are dynamic
dimension parameters; the -i
flag enables shape inference):
nnx prepare model.onnx model-prepared.onnx --set batch_size=1 --set sequence_length=255 -i
To perform shape inference programmatically, use apply_dynamic_dimensions
and infer_shapes
from the
wonnx_preprocessing::shape_inference
module.
Some models contain subgraphs whose output can be determined statically, as they do not depend on the specific inputs provided during inference. WONNX can replace such constant intermediate values with static values ('constant folding'). This is supported in the following cases:
- Output of nodes of the
Constant
op type (these are replaced with initializers) - Output of nodes of the
Shape
op type where the shape of the input is known (up front or during inference) - Output of nodes of which all inputs are constant (possibly after folding), and for which the operator is supported by WONNX.
Constant folding is performed as part of shape inference, unless disabled (from the CLI pass --no-fold-constants
to disable). This
is done in order to support models that dynamically calculate shapes using operators such as Shape
/Squeeze
/Unsqueeze
depending
on dynamically set dimension parameters (e.g. batch size).
Licensed under either of
- Apache License, Version 2.0 (LICENSE-APACHE or https://www.apache.org/licenses/LICENSE-2.0)
- MIT license (LICENSE-MIT or https://opensource.org/licenses/MIT) at your option.
Except for the following files:
-
data/models
: -
data/images
:pelican.jpeg
: source, (C) Rui Ornelas, CC-BY 2.0.bald_eagle.jpeg
: source, (C) David R. Tribble, CC BY-SA 4.0
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you shall be dual licensed as above, without any additional terms or conditions.