Web Neural Network API

Draft Community Group Report,

This version:
https://webmachinelearning.github.io/webnn/
Issue Tracking:
GitHub
Inline In Spec
Editors:
Ningxin Hu (Intel Corporation)
Chai Chaoweeraprasit (Microsoft Corporation)
Explainer:
explainer.md
Polyfill:
webnn-polyfill / webnn-samples

Abstract

This document describes a dedicated low-level API for neural network inference hardware acceleration.

Status of this document

This specification was published by the Web Machine Learning Community Group. It is not a W3C Standard nor is it on the W3C Standards Track. Please note that under the W3C Community Contributor License Agreement (CLA) there is a limited opt-out and other conditions apply. Learn more about W3C Community and Business Groups.

1. Introduction

We’re working on this section. Meanwhile, please take a look at the explainer.

2. Use cases

2.1. Application Use Cases

This section illustrates application-level use cases for neural network inference hardware acceleration. All applications in those use cases can be built on top of pre-trained deep neural network (DNN) [models].

2.1.1. Person Detection

A user opens a web-based video conferencing application, but she temporarily leaves from her room. The application is watching whether she is in front of her PC by using object detection (for example, using object detection approaches such as [SSD] or [YOLO] that use a single DNN) to detect regions in a camera input frame that include persons.

When she comes back, the application automatically detects her and notifies other online users that she is active now.

2.1.2. Semantic Segmentation

A user joins a teleconference via a web-based video conferencing application at her desk since no meeting room in her office is available. During the teleconference, she does not wish that her room and people in the background are visible. To protect the privacy of the other people and the surroundings, the application runs a machine learning model such as [DeepLabv3+] or [MaskR-CNN] to semantically split an image into segments and replaces segments that represent other people and background with another picture.

2.1.3. Skeleton Detection

A web-based video conferencing application tracks a pose of user’s skeleton by running a machine learning model, which allows for real-time human pose estimation, such as [PoseNet] to recognize her gesture and body language. When she raises her hand, her microphone is automatically unmuted and she can start speaking on the teleconference.

2.1.4. Face Recognition

There are multiple people in the conference room and they join an online meeting using a web-based video conferencing application. The application detects faces of participants by using object detection (for example, using object detection approaches such as [SSD]) and checks whether each face was present at the previous meeting or not by running a machine learning model such as [FaceNet], which verifies whether two faces would be identical or not.

2.1.5. Facial Landmark Detection

A user wants to find new glasses that beautifully fits her on an online glasses store. The online store offers web-based try-on simulator that runs a machine learning model such as Face Alignment Network [FAN] to detect facial landmarks like eyes, nose, mouth, etc. When she chooses a pair of glasses, the simulator properly renders the selected glasses on the detected position of eyes on her facial image.

2.1.6. Style Transfer

A user is looking for cosmetics on an online store and wondering which color may fit her face. The online store shows sample facial makeup images of cosmetics, and offers makeup simulator that runs a machine learning model like [ContextualLoss] or [PairedCycleGAN] to transfer the makeup style of the sample makeup image to her facial image. She can check how the selected makeup looks like on her face by the simulator.

2.1.7. Super Resolution

A web-based video conferencing is receiving a video stream from its peer, but the resolution of the video becomes lower due to network congestion. To prevent degradation of the perceived video quality, the application runs a machine learning model for super-resolution such as [SRGAN] to generate higher-resolution video frames.

2.1.8. Image Captioning

For better accessibility, a web-based presentation application provides automatic image captioning by running a machine learning model such as [im2txt] which predicts explanatory words of the presentation slides.

2.1.9. Machine Translation

Multiple people from various countries are talking via a web-based real-time text chat application. The application translates their conversation by using a machine learning model such as [GNMT] or [OpenNMT], which translates every text into different language.

2.1.10. Emotion Analysis

A user is talking to her friend via a web-based real-time text chat application, and she is wondering how the friend feels because she cannot see the friend’s face. The application analyses the friend’s emotion by using a machine learning model such as [DeepMoji], which infers emotion from input texts, and displays an emoji that represents the estimated emotion.

2.1.11. Video Summarization

A web-based video conferencing application records received video streams, and it needs to reduce recorded video data to be stored. The application generates the short version of the recorded video by using a machine learning model for video summarization such as [Video-Summarization-with-LSTM].

2.1.12. Noise Suppression

A web-based video conferencing application records received audio streams, but usually the background noise is everywhere. The application leverages real-time noise suppression using Recurrent Neural Network such as [RNNoise] for suppressing background dynamic noise like baby cry or dog barking to improve audio experiences in video conferences.

2.2. Framework Use Cases

This section collects framework-level use cases for a dedicated low-level API for neural network inference hardware acceleration. It is expected that Machine Learning frameworks will be key consumers of the Web Neural Network API (WebNN API) and the low-level details exposed through the WebNN API are abstracted out from typical web developers. However, it is also expected that web developers with specific interest and competence in Machine Learning will want to interface with the WebNN API directly instead of a higher-level ML framework.

2.2.1. Custom Layer

A web application developer wants to run a DNN model on the WebNN API. However, she has found that some of activation functions like [LeakyReLU], [ELU], etc. are not included in the WebNN API. To address this issue, she constructs custom layers of the additional activation functions on top of the WebNN API. Note that the scope of custom layers may include convolution, normalization, etc. as well as activation.

2.2.2. Network Concatenation

A web application uses a DNN model, and its model data of upper convolutional layers and lower fully-connected layers are stored in separate files, since model data of the fully-connected layers are periodically updated due to fine tuning at the server side.

Therefore, the application downloads both partial model files at first and concatenates them into a single model. When the model is updated, the application downloads fine-tuned part of the model and replace only the fully-connected layers with it.

2.2.3. Performance Adaptation

A web application developer has a concern about performance of her DNN model on mobile devices. She has confirmed that it may run too slow on mobile devices which do not have GPU acceleration. To address this issue, her web application refers to the WebNN API to confirm whether acceleration is available or not, so that the application can display the warning for devices without acceleration.

After several weeks, she has developed a tiny DNN model that can even run on CPU. In order to accommodate CPU execution, she modifies the application so that the application loads the tiny model in the case of CPU-only devices.

2.2.4. Operation Level Execution

A JavaScript ML framework is responsible for loading, interpreting and executing a ML model. During the model execution phase, the framework iterates through the operations of the model and executes each operation on the hardware device, like CPU, GPU or ML accelerator. To avoid the unnecessary data copying across devices, the framework selects the same device to execute the operations. For a compute intensive operation, such as convolution 2D or matrix multiplication, the framework uses WebNN API to execute it with the ML-specific acceleration available on that selected device.

3. Programming Model

3.1. Timelines

This section is non-normative.

A computer system with a user agent at the front-end and ML device at the back-end has components working on different timelines in parallel:

Content timeline

Associated with the execution of the Web script. It includes calling all methods described by this specification.

Steps executed on the content timeline look like this.
Device timeline

Associated with the ML device operations that are issued by the user agent. It includes creation of ML devices and resources and state objects, which are typically synchronous operations from the point of view of the user agent part that controls the ML device, but can live in a separate OS process.

Steps executed on the device timeline look like this.
Queue timeline

Associated with the execution of operations on the compute units of the ML device. It includes actual copy and compute jobs that run on the ML device.

Steps executed on the queue timeline look like this.

In this specification, asynchronous operations are used when the result value depends on work that happens on any timeline other than the Content timeline. They are represented by callbacks and promises in JavaScript.

MLGraph.compute():
  1. User issues a compute request by calling MLGraph.compute() on the Content timeline and gets a promise in return.

  2. User agent processes the compute request on the Device timeline by calling the OS ML API.

  3. After the ML device operating on Queue timeline is done, the user agent makes the results ready to be consumed by user and resolves the promise.

3.2. Device Selection

An MLContext interface represents a global state of neural network execution. One of the important context states is the underlying execution device that manages the resources and facilitates the compilation and the eventual execution of the neural network graph. An MLContext could be created from a specific GPU device such as GPUDevice or WebGLRenderingContext that is already in use by the application, in which case the corresponding GPUBuffer or WebGLBuffer resources used as graph constants, as well as the GPUTexture and WebGLTexture as graph inputs must also be created from the same device. In a multi-adapter configuration, the device used for MLContext must be created from the same adapter as the device used to allocate the resources referenced in the graph.

In a situation when a GPU context executes a graph with a constant or an input in the system memory as an ArrayBufferView, the input content is automatically uploaded from the system memory to the GPU memory, and downloaded back to the system memory of an ArrayBufferView output buffer at the end of the graph execution. This data upload and download cycles will only occur whenever the execution device requires the data to be copied out of and back into the system memory, such as in the case of the GPU. It doesn’t occur when the device is a CPU device. Additionally, the result of the graph execution is in a known layout format. While the execution may be optimized for a native memory access pattern in an intermediate result within the graph, the output of the last operation of the graph must convert the content back to a known layout format at the end of the graph in order to maintain the expected behavior from the caller’s perspective.

When an MLContext is created with MLContextOptions, the user agent selects and creates the underlying execution device by taking into account the application’s preference specified in the MLPowerPreference and the MLDevicePreference options:

The following table summarizes the types of resource supported by the device selected.

Device Type ArrayBufferView GPUBuffer GPUTexture WebGLBuffer WebGLTexture
GPUDevice Yes Yes Yes No No
WebGLRenderingContext Yes No No Yes Yes
default Yes No No No No
gpu Yes No No No No
cpu Yes No No No No

4. API

4.1. navigator.ml

A ML object is available in the Window and DedicatedWorkerGlobalScope contexts through the Navigator and WorkerNavigator interfaces respectively and is exposed via navigator.ml:

interface mixin NavigatorML {
  [SecureContext, SameObject] readonly attribute ML ml;
};
Navigator includes NavigatorML;
WorkerNavigator includes NavigatorML;

4.2. ML

enum MLDevicePreference {
  "default",
  "gpu",
  "cpu"
};

enum MLPowerPreference {
  // Let the user agent select the most suitable behavior.
  "default",

  // Prioritizes execution speed over power consumption.
  "high-performance",
  
  // Prioritizes power consumption over other considerations such as execution speed.
  "low-power"
};

dictionary MLContextOptions {
  // Preferred kind of device used
  MLDevicePreference devicePreference = "default";

  // Preference as related to power consumption
  MLPowerPreference powerPreference = "default";
};

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface ML {
  // Create a context with options
  MLContext createContext(optional MLContextOptions options = {});

  // Create a context from WebGL rendering context
  MLContext createContext(WebGLRenderingContext glContext);

  // Create a context from WebGPU device
  MLContext createContext(GPUDevice gpuDevice);
};
======= The createContext() method steps are:
  1. If the responsible document is not allowed to use the webnn feature, then throw a "SecurityError" DOMException and abort these steps.

  2. Let context be a new MLContext object.

  3. Switch on the method’s first argument:

    MLContextOptions
    Set context’s context type to default.
    WebGLRenderingContext
    Set context’s context type to webgl.
    GPUDevice
    Set context’s context type to webgpu.
    Otherwise
    Set context’s context type to default.
  4. Return context.

4.2.1. Permissions Policy Integration

This specification defines a policy-controlled feature identified by the string "webnn". Its default allowlist is 'self'.

4.3. MLContext

The MLContext interface represents a global state of neural network compute workload and execution processes.
[SecureContext, Exposed=(Window, DedicatedWorker)]
interface MLContext {};

The context type for an MLContext is either "default", "webgl" or "webgpu".

4.4. MLOperandDescriptor

enum MLInputOperandLayout {
  "nchw",
  "nhwc"
};

enum MLOperandType {
  "float32",
  "float16",
  "int32",
  "uint32",
  "int8",
  "uint8"
};

dictionary MLOperandDescriptor {
  // The operand type.
  required MLOperandType type;

  // The dimensions field is only required for tensor operands.
  // The negative value means an unknown dimension.
  sequence<long> dimensions;
};

4.5. MLOperand

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface MLOperand {};

4.6. MLGraphBuilder

The MLGraphBuilder interface defines a set of operations as identified by the § 2 Use cases that can be composed into a computational graph. It also represents the intermediate state of a graph building session.

typedef record<DOMString, MLOperand> MLNamedOperands;

dictionary MLBufferResourceView {
  required (WebGLBuffer or GPUBuffer) resource;
  unsigned long long offset = 0;
  unsigned long long size;
};

typedef (ArrayBufferView or MLBufferResourceView) MLBufferView;

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface MLGraphBuilder {
  // Construct the graph builder from the context.
  constructor(MLContext context);

  // Create an operand for a graph input.
  MLOperand input(DOMString name, MLOperandDescriptor desc);

  // Create an operand for a graph constant.
  MLOperand constant(MLOperandDescriptor desc, MLBufferView bufferView);

  // Create a single-value operand from the specified number of the specified type.
  MLOperand constant(double value, optional MLOperandType type = "float32");

  // Compile the graph up to the specified output operands
  Promise<MLGraph> build(MLNamedOperands outputs);
};

4.6.1. batchNormalization

Normalize the tensor values of input features across the batch dimension using [Batch-Normalization]. For each input feature, the mean and variance values of that feature supplied in this calculation as parameters are previously computed across the batch dimension of the input during the model training phrase of this operation.
dictionary MLBatchNormalizationOptions {
  MLOperand scale;
  MLOperand bias;
  long axis = 1;
  float epsilon = 1e-5;
};

partial interface MLGraphBuilder {
  MLOperand batchNormalization(MLOperand input, MLOperand mean, MLOperand variance,
                             optional MLBatchNormalizationOptions options = {});
};
Arguments:

Returns: an MLOperand. The batch-normalized N-D tensor of the same shape as the input tensor.

When input is a 4-D tensor of the "nchw" or "nhwc" layout, options.axis should be set to 1 or 3 respectively. The axis value designates the feature or channel count dimension of the input tensor.

The behavior of this operation when the input tensor is 4-D of the "nchw" layout can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
const shape = [1,-1,1,1];
return builder.add(
    builder.mul(
      builder.reshape(options.scale, shape),
      builder.div(
        builder.sub(input, builder.reshape(mean, shape)),
        builder.pow(
          builder.add(builder.reshape(variance, shape), builder.constant(options.epsilon)),
          builder.constant(0.5))
        )
      ),
    builder.reshape(options.bias, shape)
  );

4.6.2. clamp

Clamp the input tensor element-wise within a range specified by the minimum and maximum values.
dictionary MLClampOptions {
  MLOperand minValue;
  MLOperand maxValue;
};

partial interface MLGraphBuilder {
  MLOperand clamp(MLOperand x, optional MLClampOptions options = {});
};
Arguments:

Returns: an MLOperand. The output tensor of the same shape as x.

Clamp the input tensor element-wise within a range specified by minValue and maxValue. The calculation follows the expression min(max(x, minValue), maxValue). When minValue is not specified, the clamping is not performed on the lower limit. When maxValue is not specified, the clamping is not performed on the upper limit.

The behavior of this operation can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
if (options.minValue === undefined) {
  if (options.maxValue === undefined) {
    return x;
  } else {
    return builder.min(x, options.maxValue);
  }
} else {
  if (options.maxValue === undefined) {
    return builder.max(x, options.minValue);
  } else {
    return builder.min(builder.max(x, options.minValue), options.maxValue);
  }
}

4.6.3. concat

Concatenates the input tensors along a given axis.
partial interface MLGraphBuilder {
  MLOperand concat(sequence<MLOperand> inputs, long axis);
};
Arguments:

Returns: an MLOperand. The concatenated tensor of all the inputs along the axis. The output tensor has the same shape except on the dimension that all the inputs concatenated along. The size of that dimension is computed as the sum of all the input sizes of the same dimension.

4.6.4. conv2d

Compute a 2-D convolution given 4-D input and filter tensors
enum MLFilterOperandLayout {
  "oihw",
  "hwio",
  "ohwi",
  "ihwo"
};

enum MLAutoPad {
  "explicit",
  "same-upper",
  "same-lower"
};

dictionary MLConv2dOptions {
  sequence<long> padding;
  sequence<long> strides;
  sequence<long> dilations;
  sequence<long> outputPadding;
  sequence<long> outputSizes;
  MLAutoPad autoPad = "explicit";
  boolean transpose = false;
  long groups = 1;
  MLInputOperandLayout inputLayout = "nchw";
  MLFilterOperandLayout filterLayout = "oihw";
};

partial interface MLGraphBuilder {
  MLOperand conv2d(MLOperand input, MLOperand filter, optional MLConv2dOptions options = {});
};
Arguments:

Returns: an MLOperand. The output 4-D tensor that contains the convolution result. The output shape is interpreted according to the options.layout value. More specifically the sizes of the last two dimensions of the output tensor, the spatial dimensions, for the convolution operation can be calculated as follow:

output size = 1 + (input size - filter size + beginning padding + ending padding) / stride

Whereas for the transposed convolution case with options.transpose set to true, unless the options.outputSizes values are explicitly specified, the options.outputPadding may be needed to compute the spatial dimension values of the output tensor as follow:

output size = (input size - 1) * stride + filter size - beginning padding - ending padding + output padding

A depthwise conv2d operation is a variant of grouped convolution, used in models like the MobileNet, where the options.groups = input_channels = output_channels and the shape of filter tensor is [options.groups, 1, height, width] for "oihw" layout, [height, width, 1, options.groups] for "hwio" layout, [options.groups, height, width, 1] for "ohwi" layout and [1, height, width, options.groups] for "ihwo" layout.

4.6.5. element-wise binary operations

Compute the element-wise binary addition, subtraction, multiplication, division, maximum and minimum of the two input tensors.
partial interface MLGraphBuilder {
  MLOperand add(MLOperand a, MLOperand b);
  MLOperand sub(MLOperand a, MLOperand b);
  MLOperand mul(MLOperand a, MLOperand b);
  MLOperand div(MLOperand a, MLOperand b);
  MLOperand max(MLOperand a, MLOperand b);
  MLOperand min(MLOperand a, MLOperand b);
  MLOperand pow(MLOperand a, MLOperand b);
};
Arguments:

Returns: an MLOperand. The output tensor that contains the result of element-wise binary operation of the two input tensors.

The element-wise binary operation will be broadcasted according to [numpy-broadcasting-rule]. The rank of the output tensor is the maximum rank of the input tensors. For each dimension of the output tensor, its size is the maximum size along that dimension of the input tensors.

Operation types:

4.6.6. element-wise unary operations

Compute the element-wise unary operation for input tensor.
partial interface MLGraphBuilder {
  MLOperand abs(MLOperand x);
  MLOperand ceil(MLOperand x);
  MLOperand cos(MLOperand x);
  MLOperand exp(MLOperand x);
  MLOperand floor(MLOperand x);
  MLOperand log(MLOperand x);
  MLOperand neg(MLOperand x);
  MLOperand relu(MLOperand x);
  MLOperand sigmoid(MLOperand x);
  MLOperand sin(MLOperand x);
  MLOperand tan(MLOperand x);
  MLOperand tanh(MLOperand x);
};
Arguments:

Returns: an MLOperand. The output tensor that contains the result of element-wise unary operation of the input tensor. The shape of the output tensor is the same as the shape of input tensor.

Operation types:

4.6.7. gemm

Calculate the general matrix multiplication of the Basic Linear Algebra Subprograms. The calculation follows the expression alpha * A * B + beta * C, where A, B, and C are matrices, and A and B may optionally be transposed prior to the calculation.
dictionary MLGemmOptions {
  MLOperand c;
  float alpha = 1.0;
  float beta = 1.0;
  boolean aTranspose = false;
  boolean bTranspose = false;
};

partial interface MLGraphBuilder {
  MLOperand gemm(MLOperand a, MLOperand b, optional MLGemmOptions options = {});
};
Arguments:

Returns: an MLOperand. The output 2-D tensor that contains the calculated product of all the inputs.

The behavior of this operation can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
if (options.aTranspose)
  a = builder.transpose(a);

if (options.bTranspose)
  b = builder.transpose(b);

let ab = builder.matmul(builder.mul(builder.constant(options.alpha), a), b);
return (c ? builder.add(ab, builder.mul(builder.constant(options.beta), c)) : ab);

4.6.8. gru

Gated Recurrent Unit [GRU] recurrent network using an update gate and a reset gate to compute the hidden state that rolls into the output across the temporal sequence of the Network
enum MLRecurrentNetworkWeightLayout {
  "zrn",  // update-reset-new gate ordering
  "rzn"   // reset-update-new gate ordering
};

enum MLRecurrentNetworkActivation {
  "relu",
  "sigmoid",
  "tanh"
};

enum MLRecurrentNetworkDirection {
  "forward",
  "backward",
  "both"
};

dictionary MLGruOptions {
  MLOperand bias;
  MLOperand recurrentBias;
  MLOperand initialHiddenState;
  boolean resetAfter = true;
  boolean returnSequence = false;
  MLRecurrentNetworkDirection direction = "forward";
  MLRecurrentNetworkWeightLayout layout = "zrn";
  sequence<MLRecurrentNetworkActivation> activations;
};

partial interface MLGraphBuilder {
  sequence<MLOperand> gru(MLOperand input, MLOperand weight, MLOperand recurrentWeight, 
                        long steps, long hiddenSize, optional MLGruOptions options = {});
};
Arguments:

Returns: a sequence of MLOperand. The first element of the sequence is a 3-D tensor of shape [num_directions, batch_size, hidden_size], the cell output from the last time step of the network. Additionally, if returnSequence is set to true, the second element is the 4-D output tensor of shape [steps, num_directions, batch_size, hidden_size] containing every cell outputs from each time step in the temporal sequence.

The behavior of this operation can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
const numDirections = (options.direction == "both" ? 2 : 1);
let hiddenState = options.initialHiddenState;

if (!hiddenState) {
  const desc = { type: 'float32', dimensions: [numDirections, 1, hiddenSize] };
  const totalSize = numDirections * hiddenSize;
  hiddenState = builder.constant(desc, new Float32Array(totalSize).fill(0));
}

let sequence = null;
let cellWeight = [];
let cellRecurrentWeight = [];
let cellBias = [];
let cellRecurrentBias = [];

for (let slot = 0; slot < numDirections; ++slot) {
  cellWeight.push(builder.squeeze(builder.slice(weight, [slot, 0, 0], [1, -1, -1]), { axes: [0] }));
  cellRecurrentWeight.push(builder.squeeze(builder.slice(recurrentWeight, [slot, 0, 0], [1, -1, -1]), { axes: [0] }));
  cellBias.push(options.bias ? (builder.squeeze(builder.slice(options.bias, [slot, 0], [1, -1]), { axes: [0] })) : null);
  cellRecurrentBias.push(options.recurrentBias ? 
    (builder.squeeze(builder.slice(options.recurrentBias, [slot, 0], [1, -1]), { axes: [0] })) : null);
}

for (let step = 0; step < steps; ++step) {
  let cellHidden = [];
  let cellOutput = null;

  for (let slot = 0; slot < numDirections; ++slot) {
    cellHidden.push(builder.squeeze(builder.slice(hiddenState, [slot, 0, 0], [1, -1, -1]), { axes: [0] }));
  }

  for (let slot = 0; slot < numDirections; ++slot) {
    let slice = (slot == 1 || options.direction == "backward" ? steps - step - 1 : step);
    let cellInput = builder.squeeze(builder.slice(input, [slice, 0, 0], [1, -1, -1]), { axes: [0] });

    let result = builder.reshape(
      builder.gruCell(
        cellInput, cellWeight[slot], cellRecurrentWeight[slot],
        cellHidden[slot], hiddenSize, { bias: cellBias[slot],
        recurrentBias: cellRecurrentBias[slot], resetAfter: options.resetAfter,
        layout: options.layout, activations: options.activations }),
      [1, -1, hiddenSize]);

    cellOutput = (cellOutput ? builder.concat([cellOutput, result], 0) : result);
  }

  hiddenState = cellOutput;

  if (options.returnSequence) {
    cellOutput = builder.reshape(cellOutput, [1, numDirections, -1, hiddenSize]);
    sequence = (sequence ? builder.concat([sequence, cellOutput], 0) : cellOutput);
  }
}

return (sequence ? [hiddenState, sequence] : [hiddenState]);

4.6.9. gruCell

A single time step of the Gated Recurrent Unit [GRU] recurrent network using an update gate and a reset gate to compute the hidden state that rolls into the output across the temporal sequence of a recurrent network.
dictionary MLGruCellOptions {
  MLOperand bias;
  MLOperand recurrentBias;
  boolean resetAfter = true;
  MLRecurrentNetworkWeightLayout layout = "zrn";
  sequence<MLRecurrentNetworkActivation> activations;
};

partial interface MLGraphBuilder {
  MLOperand gruCell(MLOperand input, MLOperand weight, MLOperand recurrentWeight, 
                  MLOperand hiddenState, long hiddenSize, optional MLGruCellOptions options = {});
};
Arguments:

Returns: an MLOperand. The 2-D tensor of shape [batch_size, hidden_size], the cell output hidden state of a single time step of the recurrent network.

The behavior of this operation can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
const one = builder.constant(1);
const zero = builder.constant(0);

// update gate
let z = builder.sigmoid(
  builder.add(
    builder.add(
      (options.bias ? builder.slice(options.bias, [0], [hiddenSize]) : zero), 
      (options.recurrentBias ? builder.slice(options.recurrentBias, [0], [hiddenSize]) : zero)
      ),
    builder.add(
      builder.matmul(
        input, 
        builder.transpose(builder.slice(weight, [0, 0], [hiddenSize, -1]))
        ),
      builder.matmul(
        hiddenState,
        builder.transpose(builder.slice(recurrentWeight, [0, 0], [hiddenSize, -1]))
        )
      )
    )
  );

// reset gate
let r = builder.sigmoid(
  builder.add(
    builder.add(
      (options.bias ? builder.slice(options.bias, [hiddenSize], [hiddenSize]) : zero),
      (options.recurrentBias ? builder.slice(options.recurrentBias, [hiddenSize], [hiddenSize]) : zero)
      ),
    builder.add(
      builder.matmul(
        input, 
        builder.transpose(builder.slice(weight, [hiddenSize, 0], [hiddenSize, -1]))
        ),
      builder.matmul(
        hiddenState, 
        builder.transpose(builder.slice(recurrentWeight, [hiddenSize, 0], [hiddenSize, -1]))
        )
      )
    )
  );

// new gate
let n;
if (resetAfter) {
  n = builder.tanh(
    builder.add(
      (options.bias ? builder.slice(options.bias, [2 * hiddenSize], [hiddenSize]) : zero),
      builder.add(
        builder.matmul(
          input, 
          builder.transpose(builder.slice(weight, [2 * hiddenSize, 0], [hiddenSize, -1]))
          ),
        builder.mul(
          r,
          builder.add(
            (options.recurrentBias ? builder.slice(options.recurrentBias, [2 * hiddenSize], [hiddenSize]) : zero),
            builder.matmul(
              hiddenState, 
              builder.transpose(builder.slice(recurrentWeight, [2 * hiddenSize, 0], [hiddenSize, -1]))
              )
            )
          )
        )
      )
    );
}
else {
  n = builder.tanh(
    builder.add(
      builder.add(
        (options.bias ? builder.slice(options.bias, [2 * hiddenSize], [hiddenSize]) : zero),
        (options.recurrentBias ? builder.slice(options.recurrentBias, [2 * hiddenSize], [hiddenSize]) : zero)
        ),
      builder.add(
        builder.matmul(
          input, 
          builder.transpose(builder.slice(weight, [2 * hiddenSize, 0], [hiddenSize, -1]))
          ),
        builder.matmul(
          builder.mul(r, hiddenState),
          builder.transpose(builder.slice(recurrentWeight, [2 * hiddenSize, 0], [hiddenSize, -1]))
          )
        )
      )
    );
}

// compute the new hidden state
return builder.add(builder.mul(z, hiddenState), builder.mul(n, builder.sub(one, z)));

4.6.10. instanceNormalization

Normalize the input features using [Instance-Normalization]. Unlike § 4.6.1 batchNormalization where the mean and variance values used in the calculation are previously computed across the batch dimension during the model training phrase, the mean and variance values used in the calculation of an instance normalization are computed internally on the fly per input feature.
dictionary MLInstanceNormalizationOptions {
  MLOperand scale;
  MLOperand bias;
  float epsilon = 1e-5;
  MLInputOperandLayout layout = "nchw";
};

partial interface MLGraphBuilder {
  MLOperand instanceNormalization(MLOperand input, 
                                optional MLInstanceNormalizationOptions options = {});
};
Arguments:

Returns: an MLOperand. The instance-normalized 4-D tensor of the same shape as the input tensor.

The behavior of this operation when the input tensor is 4-D of the "nchw" layout can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
// The mean reductions happen over the spatial dimensions of the input
// e.g. axis 2 and 3 of the input tensor.
const reduceOptions = { axes: [2,3], keepDimensions: true };
const mean = builder.reduceMean(input, reduceOptions);
const variance = builder.reduceMean(
  builder.pow(
    builder.sub(input, mean), 
    buider.constant(2)),
  reduceOptions
  );

// The scale and bias values are applied per input feature
// e.g. axis 1 of the input tensor.
const shape = [1,-1,1,1];
return builder.add(
  builder.mul(
    builder.reshape(options.scale, shape),
    builder.div(
      builder.sub(input, mean),
      buidler.pow(
        builder.add(variance, options.epsilon), 
        builder.constant(0.5))
      )
    ),
  builder.reshape(options.bias, shape)
  );

4.6.11. leakyRelu

dictionary MLLeakyReluOptions {
  float alpha = 0.01;
};

partial interface MLGraphBuilder {
  MLOperand leakyRelu(MLOperand x, optional MLLeakyReluOptions options = {});
};
Arguments:

Returns: an MLOperand. The output tensor of the same shape as x.

Calculate the leaky version of rectified linear function on the input tensor element-wise. The calculation follows the expression max(0, x) + alpha ∗ min(0, x).

The behavior of this operation can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
return builder.add(builder.max(builder.constant(0), x),
          builder.mul(builder.constant(options.alpha), builder.min(builder.constant(0), x)));

4.6.12. matmul

Compute the matrix product of two input tensors.
partial interface MLGraphBuilder {
  MLOperand matmul(MLOperand a, MLOperand b);
};
Arguments:

Returns: an MLOperand. The output N-D tensor that contains the matrix product of two input tensors.

Compute the matrix product of two input tensors. It behaves as following:

4.6.13. pad

Inflate the tensor with constant or mirrored values on the edges.
enum MLPaddingMode {
  "constant",
  "edge",
  "reflection",
  "symmetric"
};

dictionary MLPadOptions {
  MLPaddingMode mode = "constant";
  float value = 0;
};

partial interface MLGraphBuilder {
  MLOperand pad(MLOperand input, MLOperand padding, optional MLPadOptions options = {});
};
Arguments:

Returns: an MLOperand. The padded output tensor.

// input: [[1,2,3], [4,5,6]]
const input = builder.constant(
  { type: 'float32', dimensions: [2,3] }, new Float32Array([1,2,3,4,5,6]));

// padding: [[1,1], [2,2]]
const padding = builder.constant(
  { type: 'float32', dimensions: [2,2] }, new Float32Array([1,1,2,2]));

// "constant" padded:
//    [[0,0,0,0,0,0,0],
//     [0,0,1,2,3,0,0],
//     [0,0,4,5,6,0,0],
//     [0,0,0,0,0,0,0]]
builder.pad(input, padding);

// "edge" padded:
//    [[1,1,1,2,3,3,3],
//     [1,1,1,2,3,3,3],
//     [4,4,4,5,6,6,6],
//     [4,4,4,5,6,6,6]]
builder.pad(input, padding, { mode: "edge" });

// "reflection" padded:
//    [[6,5,4,5,6,5,4],
//     [3,2,1,2,3,2,1],
//     [6,5,4,5,6,5,4],
//     [3,2,1,2,3,2,1]]
builder.pad(input, padding, { mode: "reflection" });

// "symmetric" padded:
//    [[2,1,1,2,3,3,2],
//     [2,1,1,2,3,3,2],
//     [5,4,4,5,6,6,5],
//     [5,4,4,5,6,6,5]]
builder.pad(input, padding, { mode: "symmetric" });

4.6.14. pooling operations

Compute a mean, L2 norm, or max reduction operation across all the elements within the moving window over the input tensor. See the description of each type of reduction in § 4.6.15 reduction operations.
dictionary MLPool2dOptions {
  sequence<long> windowDimensions;
  sequence<long> padding;
  sequence<long> strides;
  sequence<long> dilations;
  MLAutoPad autoPad = "explicit";
  MLInputOperandLayout layout = "nchw";
};

partial interface MLGraphBuilder {
  MLOperand averagePool2d(MLOperand input, optional MLPool2dOptions options = {});
  MLOperand l2Pool2d(MLOperand input, optional MLPool2dOptions options = {});
  MLOperand maxPool2d(MLOperand input, optional MLPool2dOptions options = {});
};
Arguments:

Returns: an MLOperand. The output 4-D tensor that contains the result of the reduction. The logical shape is interpreted according to the value of layout.

A global pooling operation such as one for the max pooling operation is a variant of pooling where the window dimensions is the spatial dimensions (last two dimensions) of the input shape, as follow.
// 'global' max pooling
builder.maxPool2d(input);

4.6.15. reduction operations

Reduce the input along the dimensions given in axes.
dictionary MLReduceOptions {
  sequence<long> axes = null;
  boolean keepDimensions = false;
};

partial interface MLGraphBuilder {
  MLOperand reduceL1(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceL2(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceLogSum(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceLogSumExp(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceMax(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceMean(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceMin(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceProduct(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceSum(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceSumSquare(MLOperand input, optional MLReduceOptions options = {});
};
Arguments:

Returns: an MLOperand. The reduced output tensor.

Reduction types:

4.6.16. resample

Resample the tensor values from the source to the destination dimensions according to the scaling factors.
enum MLInterpolationMode {
  "nearest-neighbor",
  "linear"
};

dictionary MLResampleOptions {
  MLInterpolationMode mode = "nearest-neighbor";
  sequence<float> scales;
  sequence<long> sizes;
};

partial interface MLGraphBuilder {
  MLOperand resample(MLOperand input, optional MLResampleOptions options = {});
};
Arguments:

Returns: an MLOperand. The output 4-D tensor.

4.6.17. reshape

Alter the shape of a tensor to a new shape. Reshape does not copy or change the content of the tensor. It just changes the tensor’s logical dimensions for the subsequent operations.
partial interface MLGraphBuilder {
  MLOperand reshape(MLOperand input, sequence<long> newShape);
};
Arguments:

Returns: an MLOperand. The output tensor. The values of the output tensor are the same as values of the input tensor. The shape of the output tensor is specified by the newShape argument.

4.6.18. slice

Produce a slice of the input tensor.
dictionary MLSliceOptions {
  sequence<long> axes;
};

partial interface MLGraphBuilder {
  MLOperand slice(MLOperand input, sequence<long> starts, sequence<long> sizes,
                optional MLSliceOptions options = {});
};
Arguments:

Returns: an MLOperand. The output tensor of the same rank as the input tensor with tensor values stripped to the specified starting and ending indices in each dimension.

4.6.19. softmax

Compute the softmax values of the 2-D input tensor along axis 1.
partial interface MLGraphBuilder {
  MLOperand softmax(MLOperand x);
};
Arguments:

Returns: an MLOperand. The output 2-D tensor that contains the softmax results, of the same shape as the input tensor.

The behavior of this operation can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
// This sample deploys a well-known implementation trick [1] to compute the
// exponentials of the distances to the max value, instead of the exponentials
// of the input values itself, in order to increase the numerical stability of
// the result.
// [1]: https://cs231n.github.io/linear-classify/#softmax
const max_x = builder.reduceMax(x, { axes: [1], keepDimensions: true });
const exp_x = builder.exp(builder.sub(x, max));
return builder.div(exp_x, builder.reduceSum(exp_x, { axes: [1], keepDimensions: true }));

4.6.20. split

Split the input tensor into a number of sub tensors along the given axis.
dictionary MLSplitOptions {
  long axis = 0;
};

partial interface MLGraphBuilder {
  sequence<MLOperand> split(MLOperand input,
                          (unsigned long or sequence<unsigned long>) splits,
                          optional MLSplitOptions options = {});
};
Arguments:

Returns: a sequence of MLOperand. The splitted output tensors. If splits is an unsigned long, the length of the output sequence equals to splits. The shape of each output tensor is the same as input except the dimension size of axis equals to the quotient of dividing the dimension size of input along axis by splits. If splits is a sequence of unsigned long, the length of the output sequence equals to the length of splits. The shape of the i-th output tensor is the same as as input except along axis where the dimension size is splits[i].

The behavior of this operation can be generically emulated from the usage of other operations as follow. However, user agents typically have a more efficient implementation for it, therefore its usage is encouraged from the performance standpoint.
// This sample shows the case that the splits parameter is an array.
const outputs = [];
let start = 0;
for (const size of splits) {
  outputs.push(builder.slice(input, [start], [size], { axis: [options.axis] }));
  start += size;
}
return outputs;

4.6.21. squeeze

Reduce the rank of a tensor by eliminating dimensions with size 1 of the tensor shape. Squeeze only affects the tensor’s logical dimensions. It does not copy or change the content in the tensor.
dictionary MLSqueezeOptions {
  sequence<long> axes;
};

partial interface MLGraphBuilder {
  MLOperand squeeze(MLOperand input, optional MLSqueezeOptions options = {});
};
Arguments:

Returns: an MLOperand. The output tensor of the same or reduced rank with the shape dimensions of size 1 eliminated.

4.6.22. transpose

Permute the dimensions of the input tensor according to the permutation argument.
dictionary MLTransposeOptions {
  sequence<long> permutation;
};

partial interface MLGraphBuilder {
  MLOperand transpose(MLOperand input, optional MLTransposeOptions options = {});
};
Arguments:

Returns: an MLOperand. The permuted or transposed N-D tensor.

4.7. MLGraph

The MLGraph interface represents a compiled computational graph. A compiled graph once constructed is immutable and cannot be subsequently changed.
dictionary MLInput {
  required (MLBufferView or WebGLTexture or GPUTexture) data;
  sequence<long> dimensions;
};

dictionary MLOutput {
  (MLBufferView or WebGLTexture or GPUTexture) data;
  sequence<long> dimensions;
};

typedef record<DOMString, MLInput> MLNamedInputs;
typedef record<DOMString, MLOutput> MLNamedOutputs;

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface MLGraph {
  Promise<MLNamedOutputs> compute(MLNamedInputs inputs, 
                                  optional MLNamedOutputs outputs = {});
};

MLGraph has the following internal slots:

[[context]] of type MLContext

The context of type MLContext associated with this MLGraph.

[[inputOperands]] of type record<DOMString, MLOperandDescriptor>

Maps the name of an input MLOperand to its MLOperandDescriptor for all input MLOperands of this MLGraph.

[[outputOperands]] of type sequence<DOMString>

Contains the names of all output MLOperands of this MLGraph.

[[implementation]]

The underlying implemenation provided by the User Agent.

compute(inputs, outputs)

Issue a compute request of the MLGraph given MLNamedInputs and optional MLNamedOutputs. The returned Promise resolves when the results in MLNamedOutputs are ready to be consumed.

Called on: MLGraph this.

Arguments:

Arguments for the MLGraph.compute(inputs, outputs) method.
Parameter Type Nullable Optional Description
inputs MLNamedInputs a MLNamedInputs. The data and optional dimensions of inputs for the compute request.
outputs MLNamedOutputs an optional MLNamedOutputs. The names and pre-allocated resources of required outputs for the compute request. Default to be an empty record which means that the compute request is for all outputs.

Returns: Promise<MLNamedOutputs>. The dimensions and data of outputs returned by the compute request.

  1. Let promise be a new promise.

  2. If any of the following requirements are unmet, then reject promise with a TypeError and stop.

    1. For each key -> value of inputs:

      1. this.[[inputOperands]][key] must exist.

      2. Let inputOperand be this.[[inputOperands]][key].

      3. If value.data is an ArrayBufferView, then:

        1. The kind of value.data must be compatible to inputOperand.type according to this table.

      4. If value.dimensions was given, then:

        1. The length of value.dimensions must be the same as the length of inputOperand.dimensions.

        2. Let i be 0.

        3. While true:

          1. Let dimension be value.dimensions[i].

          2. dimension must be greater than 0.

          3. If inputOperand.dimensions[i] is greater than 0, then dimension must be equal to inputOperand.dimensions[i].

          4. Set i to i + 1.

          5. If i if equal to the length of value.dimensions, then break.

      5. Else:

        1. For each dimension of inputOperand.dimensions:

          1. The value of dimension must be greater than 0.

    2. If outputs was not an empty record, then:

      1. For each key -> value of outputs:

        1. this.[[outputOperands]][key] must exist.

        2. If value.data was given, then the kind of value.data must be compatible to this.[[outputOperands]][key] according to this table.

  3. Let requiredOutputNames be a new ordered set<DOMString>.

  4. If outputs was not an empty record, then:

    1. For each key -> value of outputs:

      1. Append key to requiredOutputNames.

  5. Else:

    1. For each key -> value of this.[[outputOperands]]:

      1. Append key to requiredOutputNames.

  6. Let copiedInputs be a new MLNamedInputs.

  7. For each key -> value of inputs:

    1. Let copiedInputs be a new MLInput.

    2. Let copiedInputs.data be a new ArrayBufferView that has the same kind and length as value.data's.

    3. Set the content of copiedInputs.data to the content of value.data.

    4. Let copiedInputs.dimensions be a new sequence<long> that has the same length of value.dimensions's.

    5. Set the content of copiedInputs.dimensions to the content of value.dimensions.

    6. Set copiedInputs[key] to copiedInputs.

  8. Let results be a new MLNamedOutputs.

  9. Let remainingOutputNames be a new ordered set<DOMString>.

  10. Set the content of remainingOutputNames to the content of requiredOutputNames.

  11. Issue the following steps on the Device timeline of this.[[implementation]]:

    1. For each outputName of requiredOutputNames:

      1. Issue a compute request of this.[[implementation]] for output whose name is outputName with given copiedInputs.

      2. When the compute request is completed, issue the following steps on the appropriate Queue timeline:

        1. If there is an error returned by this.[[implementation]], then:

          1. reject promise with an OperationError and stop.

        2. Else:

          1. Let outputRank be a unsigned long.

          2. Set outputRank to the rank of output tensor returned by this.[[implementation]].

          3. Let outputDemisions be a new sequence<long> of size outputRank.

          4. Let i be 0.

          5. Let outputSize to 1.

          6. While true:

            1. Set outputDimensions[i] to the dimension at ith axis of output tensor returned by this.[[implementation]].

            2. Set outputSize to outputSize * outputDimensions[i].

            3. Set i to i + 1.

            4. If i is equal to outputRank, then break.

          7. Set results[outputName].dimensions to outputDemisions.

          8. If this.[[context]] is created from MLContextOptions, then:

            1. If outputs[outputName].data was given, then:

              1. If outputs|[outputName].data is not an ArrayBufferView, then reject promise with an TypeError and stop.

              2. If the kind of outputs[outputName].data is not compatible to output tensor according to this table, then reject promise with a TypeError and stop.

              3. If the length of outputs[outputName].data is less than outputSize, then reject promise with a TypeError and stop.

              4. Set the content of outputs[outputName].data to the content of output tensor returned by this.[[implementation]].

            2. Else:

              1. Let results[outputName].data be a new ArrayBufferView of size outputSize and kind that is compatible to output tensor according to this table.

              2. Set the content of results[outputName].data to the content of output tensor returned by this.[[implementation]].

          9. Remove outputName from remainingOutputNames.

          10. If remainingOutputNames is empty, then resolve promise with results and stop.

  12. Return promise.

Describe the algorithm steps for this.[[context]] created from WebGLRenderingContext and GPUDevice.

4.7.1. Examples

The following code showcases the computation with dynamic input dimensions.
const context = navigator.ml.createContext();

// Create a graph with dynamic shaped inputs.
const builder = new MLGraphBuilder(context);
const descA = {type: 'float32', dimensions: [-1, 4]};
const a = builder.input('a', descA);
const descB = {type: 'float32', dimensions: [4, -1]};
const b = builder.input('b', descB);
const c = builder.matmul(a, b);
const graph = await builder.build({c});

async function compute(shapeA, shapeB) {
  const bufferA = new Float32Array(sizeOfShape(shapeA)).fill(0.5);
  const bufferB = new Float32Array(sizeOfShape(shapeB)).fill(0.5);

  // Specify the shape of inputs when computing.
  const inputs = {
    'a': {data: bufferA, dimensions: shapeA},
    'b': {data: bufferB, dimensions: shapeB},
  };
  const outputs = await graph.compute(inputs);
  console.log(`shape: [${outputs.c.dimensions}], values: ${outputs.c.data}`);
}

await compute([3, 4], [4, 3]);
await compute([4, 4], [4, 4]);
await compute([5, 4], [4, 5]);
The following code showcases the computation with pre-allocated output buffers.
const context = navigator.ml.createContext();

// The following code multiplies matrix a of shape [3, 4] with matrix b of shape [4, 3]
// into matrix c of shape [3, 3].
const builder = new MLGraphBuilder(context);
const descA = {type: 'float32', dimensions: [3, 4]};
const a = builder.input('a', descA);
const descB = {type: 'float32', dimensions: [4, 3]};
const bufferB = new Float32Array(sizeOfShape(descB.dimensions)).fill(0.5);
const b = builder.constant(descB, bufferB);
const c = builder.matmul(a, b);
const graph = await builder.build({c});

const bufferA = new Float32Array(sizeOfShape(descA.dimensions)).fill(0.5);
const inputs = {'a': {data: bufferA}};
// Pre-allocate output buffer for c.
const outputs = {'c': {data: new Float32Array(sizeOfShape([3, 3]))}};
await graph.compute(inputs, outputs);
console.log(`values: ${outputs.c.data}`);
The following code showcases the computation with optional outputs.
const context = navigator.ml.createContext();

// Build a graph with two outputs.
const builder = new MLGraphBuilder(context);
const descA = {type: 'float32', dimensions: [3, 4]};
const a = builder.input('a', descA);
const descB = {type: 'float32', dimensions: [4, 3]};
const bufferB = new Float32Array(sizeOfShape(descB.dimensions)).fill(0.5);
const b = builder.constant(descB, bufferB);
const descC = {type: 'float32', dimensions: [3, 3]};
const bufferC = new Float32Array(sizeOfShape(descC.dimensions)).fill(1);
const c = builder.constant(descC, bufferC);
const d = builder.matmul(a, b);
const e = builder.add(d, c);
const graph = await builder.build({d, e});

const bufferA = new Float32Array(sizeOfShape(descA.dimensions)).fill(0.5);
const inputs = {'a': {data: bufferA}};

// Compute both d and e.
let outputs = await graph.compute(inputs);
console.log(`outputs include ${Object.keys(outputs)}`);

// Compute d.
outputs = await graph.compute(inputs, {d});
console.log(`outputs include ${Object.keys(outputs)}`);
console.log(`shape: [${outputs.d.dimensions}], values: ${outputs.d.data}`);

// Compute e.
outputs = await graph.compute(inputs, {e});
console.log(`outputs include ${Object.keys(outputs)}`);
console.log(`shape: [${outputs.e.dimensions}], values: ${outputs.e.data}`);

5. Examples

The following code gets the MLContext object.
const context = navigator.ml.createContext({powerPreference: 'low-power'});
The following code builds a graph as:
constant1 ---+
             +--- Add ---> intermediateOutput1 ---+
input1    ---+                                    |
                                                  +--- Mul---> output
constant2 ---+                                    |
             +--- Add ---> intermediateOutput2 ---+
input2    ---+
// Use tensors in 4 dimensions.
const TENSOR_DIMS = [1, 2, 2, 2];
const TENSOR_SIZE = 8;

const builder = new MLGraphBuilder(context);

// Create MLOperandDescriptor object.
const desc = {type: 'float32', dimensions: TENSOR_DIMS};

// constant1 is a constant MLOperand with the value 0.5.
const constantBuffer1 = new Float32Array(TENSOR_SIZE).fill(0.5);
const constant1 = builder.constant(desc, constantBuffer1);

// input1 is one of the input MLOperands. Its value will be set before execution.
const input1 = builder.input('input1', desc);

// constant2 is another constant MLOperand with the value 0.5.
const constantBuffer2 = new Float32Array(TENSOR_SIZE).fill(0.5);
const constant2 = builder.constant(desc, constantBuffer2);

// input2 is another input MLOperand. Its value will be set before execution.
const input2 = builder.input('input2', desc);

// intermediateOutput1 is the output of the first Add operation.
const intermediateOutput1 = builder.add(constant1, input1);

// intermediateOutput2 is the output of the second Add operation.
const intermediateOutput2 = builder.add(constant2, input2);

// output is the output MLOperand of the Mul operation.
const output = builder.mul(intermediateOutput1, intermediateOutput2);
Compile the graph up to the output operand.
// Compile the constructed graph.
const graph = await builder.build({'output': output});
The following code executes the compiled graph.
// Setup the input buffers with value 1.
const inputBuffer1 = new Float32Array(TENSOR_SIZE).fill(1);
const inputBuffer2 = new Float32Array(TENSOR_SIZE).fill(1);

// Asynchronously execute the compiled graph with the specified inputs.
const inputs = {
  'input1': {data: inputBuffer1},
  'input2': {data: inputBuffer2},
};
const outputs = await graph.compute(inputs);

// Log the shape and computed result of the output operand.
console.log('Output shape: ' + outputs.output.dimensions);
// Output shape: 1,2,2,2
console.log('Output value: ' + outputs.output.data);
// Output value: 2.25,2.25,2.25,2.25,2.25,2.25,2.25,2.25

6. Appendices

6.1. MLOperandType and ArrayBufferView compatibility

MLOperandType ArrayBufferView
float32 Float32Array
int32 Int32Array
uint32 Uint32Array
int8 Int8Array
uint8 Uint8Array

clarify the usage of ArrayBufferView for float16. <https://github.com/webmachinelearning/webnn/issues/127>

7. Acknowledgements

This specification follows the concepts of the Android Neural Networks API C API.

Thanks to Tomoyuki Shimizu, Ningxin Hu, Zhiqiang Yu and Belem Zhang for the use cases.

Thanks to Nikhil Thorat, Daniel Smilkov, Ganesan Ramalingam, Rafael Cintron and Benjamin Poulain for their contributions to the API specification.

Conformance

Document conventions

Conformance requirements are expressed with a combination of descriptive assertions and RFC 2119 terminology. The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “MAY”, and “OPTIONAL” in the normative parts of this document are to be interpreted as described in RFC 2119. However, for readability, these words do not appear in all uppercase letters in this specification.

All of the text of this specification is normative except sections explicitly marked as non-normative, examples, and notes. [RFC2119]

Examples in this specification are introduced with the words “for example” or are set apart from the normative text with class="example", like this:

This is an example of an informative example.

Informative notes begin with the word “Note” and are set apart from the normative text with class="note", like this:

Note, this is an informative note.

Conformant Algorithms

Requirements phrased in the imperative as part of algorithms (such as "strip any leading space characters" or "return false and abort these steps") are to be interpreted with the meaning of the key word ("must", "should", "may", etc) used in introducing the algorithm.

Conformance requirements phrased as algorithms or specific steps can be implemented in any manner, so long as the end result is equivalent. In particular, the algorithms defined in this specification are intended to be easy to understand and are not intended to be performant. Implementers are encouraged to optimize.

Index

Terms defined by this specification

Terms defined by reference

References

Normative References

[HTML]
Anne van Kesteren; et al. HTML Standard. Living Standard. URL: https://html.spec.whatwg.org/multipage/
[INFRA]
Anne van Kesteren; Domenic Denicola. Infra Standard. Living Standard. URL: https://infra.spec.whatwg.org/
[NUMPY-BROADCASTING-RULE]
The SciPy community. General Broadcasting Rules of NumPy. July 2019. URL: https://docs.scipy.org/doc/numpy/user/basics.broadcasting.html#general-broadcasting-rules
[PERMISSIONS-POLICY-1]
Ian Clelland. Permissions Policy. 16 July 2020. WD. URL: https://www.w3.org/TR/permissions-policy-1/
[RFC2119]
S. Bradner. Key words for use in RFCs to Indicate Requirement Levels. March 1997. Best Current Practice. URL: https://tools.ietf.org/html/rfc2119
[WEBGL-1]
Dean Jackson; Jeff Gilbert. WebGL Specification, Version 1.0. 9 August 2017. URL: https://www.khronos.org/registry/webgl/specs/latest/1.0/
[WEBGPU]
Dzmitry Malyshau; Kai Ninomiya. WebGPU. ED. URL: https://gpuweb.github.io/gpuweb/
[WebIDL]
Boris Zbarsky. Web IDL. 15 December 2016. ED. URL: https://heycam.github.io/webidl/

Informative References

[Batch-Normalization]
Sergey Ioffe; Christian Szegedy. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. March 2015. URL: https://arxiv.org/abs/1502.03167
[ContextualLoss]
Roey Mechrez; Itamar Talmi; Lihi Zelnik-Manor. The Contextual Loss for Image Transformation with Non-Aligned Data. July 2018. URL: https://arxiv.org/abs/1803.02077
[DeepLabv3+]
Liang-Chieh Chen; et al. Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation. August 2018. URL: https://arxiv.org/abs/1802.02611
[DeepMoji]
Bjarke Felbo; et al. Using millions of emoji occurrences to learn any-domain representations for detecting sentiment, emotion and sarcasm. October 2017. URL: https://arxiv.org/abs/1708.00524
[ELU]
Djork-Arné Clevert; Thomas Unterthiner; Sepp Hochreiter. Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs). February 2016. URL: https://arxiv.org/abs/1511.07289
[FaceNet]
Florian Schroff; Dmitry Kalenichenko; James Philbin. FaceNet: A Unified Embedding for Face Recognition and Clustering. June 2015. URL: https://arxiv.org/abs/1503.03832
[FAN]
Adrian Bulat; Georgios Tzimiropoulos. How far are we from solving the 2D & 3D Face Alignment problem? (and a dataset of 230,000 3D facial landmarks). September 2017. URL: https://arxiv.org/abs/1703.07332
[GNMT]
Minh-Thang Luong; Eugene Brevdo; Rui Zhao. Neural Machine Translation (seq2seq) Tutorial. May 2017. URL: https://github.com/tensorflow/nmt
[GRU]
Kyunghyun Cho; et al. Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation. September 2014. URL: https://arxiv.org/pdf/1406.1078.pdf
[IM2TXT]
Oriol Vinyals; et al. Show and Tell: Lessons learned from the 2015 MSCOCO Image Captioning Challenge. September 2016. URL: https://arxiv.org/abs/1609.06647
[Instance-Normalization]
Dmitry Ulyanov; Andrea Vedaldi; Victor Lempitsky. Instance Normalization: The Missing Ingredient for Fast Stylization. July 2016. URL: https://arxiv.org/abs/1607.08022
[LeakyReLU]
Andrew L. Maas; Awni Y. Hannun; Andrew Y. Ng. Rectifier Nonlinearities Improve Neural Network Acoustic Models. June 2013. URL: https://pdfs.semanticscholar.org/367f/2c63a6f6a10b3b64b8729d601e69337ee3cc.pdf
[MaskR-CNN]
Kaiming He; et al. Mask R-CNN. January 2018. URL: https://arxiv.org/abs/1703.06870
[MODELS]
Machine Learning for the Web Community Group. The first-wave models. 2020. URL: https://github.com/webmachinelearning/webnn/blob/master/op_compatibility/first_wave_models.md
[OpenNMT]
Guillaume Klein; et al. OpenNMT: Open-Source Toolkit for Neural Machine Translation. March 2017. URL: https://arxiv.org/abs/1701.02810
[PairedCycleGAN]
Huiwen Chang; et al. PairedCycleGAN: Asymmetric Style Transfer for Applying and Removing Makeup. June 2018. URL: http://openaccess.thecvf.com/content_cvpr_2018/html/Chang_PairedCycleGAN_Asymmetric_Style_CVPR_2018_paper.html
[PoseNet]
Dan Oved. Real-time Human Pose Estimation in the Browser with TensorFlow.js. May 2018. URL: https://medium.com/tensorflow/real-time-human-pose-estimation-in-the-browser-with-tensorflow-js-7dd0bc881cd5
[RNNoise]
Jean-Marc Valin. Recurrent neural network for audio noise reduction. September 2017. URL: https://github.com/xiph/rnnoise
[SRGAN]
Christian Ledig; et al. Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network. May 2017. URL: https://arxiv.org/abs/1609.04802
[SSD]
Wei Liu; et al. SSD: Single Shot MultiBox Detector. December 2016. URL: https://arxiv.org/abs/1512.02325
[Video-Summarization-with-LSTM]
Ke Zhang; et al. Video summarization with long short-term memory. October 2016. URL: http://www-scf.usc.edu/~zhan355/ke_eccv2016.pdf
[YOLO]
Joseph Redmon; et al. You Only Look Once: Unified, Real-Time Object Detection. May 2016. URL: https://arxiv.org/abs/1506.02640

IDL Index

interface mixin NavigatorML {
  [SecureContext, SameObject] readonly attribute ML ml;
};
Navigator includes NavigatorML;
WorkerNavigator includes NavigatorML;

enum MLDevicePreference {
  "default",
  "gpu",
  "cpu"
};

enum MLPowerPreference {
  // Let the user agent select the most suitable behavior.
  "default",

  // Prioritizes execution speed over power consumption.
  "high-performance",
  
  // Prioritizes power consumption over other considerations such as execution speed.
  "low-power"
};

dictionary MLContextOptions {
  // Preferred kind of device used
  MLDevicePreference devicePreference = "default";

  // Preference as related to power consumption
  MLPowerPreference powerPreference = "default";
};

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface ML {
  // Create a context with options
  MLContext createContext(optional MLContextOptions options = {});

  // Create a context from WebGL rendering context
  MLContext createContext(WebGLRenderingContext glContext);

  // Create a context from WebGPU device
  MLContext createContext(GPUDevice gpuDevice);
};

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface MLContext {};

enum MLInputOperandLayout {
  "nchw",
  "nhwc"
};

enum MLOperandType {
  "float32",
  "float16",
  "int32",
  "uint32",
  "int8",
  "uint8"
};

dictionary MLOperandDescriptor {
  // The operand type.
  required MLOperandType type;

  // The dimensions field is only required for tensor operands.
  // The negative value means an unknown dimension.
  sequence<long> dimensions;
};

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface MLOperand {};

typedef record<DOMString, MLOperand> MLNamedOperands;

dictionary MLBufferResourceView {
  required (WebGLBuffer or GPUBuffer) resource;
  unsigned long long offset = 0;
  unsigned long long size;
};

typedef (ArrayBufferView or MLBufferResourceView) MLBufferView;

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface MLGraphBuilder {
  // Construct the graph builder from the context.
  constructor(MLContext context);

  // Create an operand for a graph input.
  MLOperand input(DOMString name, MLOperandDescriptor desc);

  // Create an operand for a graph constant.
  MLOperand constant(MLOperandDescriptor desc, MLBufferView bufferView);

  // Create a single-value operand from the specified number of the specified type.
  MLOperand constant(double value, optional MLOperandType type = "float32");

  // Compile the graph up to the specified output operands
  Promise<MLGraph> build(MLNamedOperands outputs);
};

dictionary MLBatchNormalizationOptions {
  MLOperand scale;
  MLOperand bias;
  long axis = 1;
  float epsilon = 1e-5;
};

partial interface MLGraphBuilder {
  MLOperand batchNormalization(MLOperand input, MLOperand mean, MLOperand variance,
                             optional MLBatchNormalizationOptions options = {});
};

dictionary MLClampOptions {
  MLOperand minValue;
  MLOperand maxValue;
};

partial interface MLGraphBuilder {
  MLOperand clamp(MLOperand x, optional MLClampOptions options = {});
};

partial interface MLGraphBuilder {
  MLOperand concat(sequence<MLOperand> inputs, long axis);
};

enum MLFilterOperandLayout {
  "oihw",
  "hwio",
  "ohwi",
  "ihwo"
};

enum MLAutoPad {
  "explicit",
  "same-upper",
  "same-lower"
};

dictionary MLConv2dOptions {
  sequence<long> padding;
  sequence<long> strides;
  sequence<long> dilations;
  sequence<long> outputPadding;
  sequence<long> outputSizes;
  MLAutoPad autoPad = "explicit";
  boolean transpose = false;
  long groups = 1;
  MLInputOperandLayout inputLayout = "nchw";
  MLFilterOperandLayout filterLayout = "oihw";
};

partial interface MLGraphBuilder {
  MLOperand conv2d(MLOperand input, MLOperand filter, optional MLConv2dOptions options = {});
};

partial interface MLGraphBuilder {
  MLOperand add(MLOperand a, MLOperand b);
  MLOperand sub(MLOperand a, MLOperand b);
  MLOperand mul(MLOperand a, MLOperand b);
  MLOperand div(MLOperand a, MLOperand b);
  MLOperand max(MLOperand a, MLOperand b);
  MLOperand min(MLOperand a, MLOperand b);
  MLOperand pow(MLOperand a, MLOperand b);
};

partial interface MLGraphBuilder {
  MLOperand abs(MLOperand x);
  MLOperand ceil(MLOperand x);
  MLOperand cos(MLOperand x);
  MLOperand exp(MLOperand x);
  MLOperand floor(MLOperand x);
  MLOperand log(MLOperand x);
  MLOperand neg(MLOperand x);
  MLOperand relu(MLOperand x);
  MLOperand sigmoid(MLOperand x);
  MLOperand sin(MLOperand x);
  MLOperand tan(MLOperand x);
  MLOperand tanh(MLOperand x);
};

dictionary MLGemmOptions {
  MLOperand c;
  float alpha = 1.0;
  float beta = 1.0;
  boolean aTranspose = false;
  boolean bTranspose = false;
};

partial interface MLGraphBuilder {
  MLOperand gemm(MLOperand a, MLOperand b, optional MLGemmOptions options = {});
};

enum MLRecurrentNetworkWeightLayout {
  "zrn",  // update-reset-new gate ordering
  "rzn"   // reset-update-new gate ordering
};

enum MLRecurrentNetworkActivation {
  "relu",
  "sigmoid",
  "tanh"
};

enum MLRecurrentNetworkDirection {
  "forward",
  "backward",
  "both"
};

dictionary MLGruOptions {
  MLOperand bias;
  MLOperand recurrentBias;
  MLOperand initialHiddenState;
  boolean resetAfter = true;
  boolean returnSequence = false;
  MLRecurrentNetworkDirection direction = "forward";
  MLRecurrentNetworkWeightLayout layout = "zrn";
  sequence<MLRecurrentNetworkActivation> activations;
};

partial interface MLGraphBuilder {
  sequence<MLOperand> gru(MLOperand input, MLOperand weight, MLOperand recurrentWeight, 
                        long steps, long hiddenSize, optional MLGruOptions options = {});
};

dictionary MLGruCellOptions {
  MLOperand bias;
  MLOperand recurrentBias;
  boolean resetAfter = true;
  MLRecurrentNetworkWeightLayout layout = "zrn";
  sequence<MLRecurrentNetworkActivation> activations;
};

partial interface MLGraphBuilder {
  MLOperand gruCell(MLOperand input, MLOperand weight, MLOperand recurrentWeight, 
                  MLOperand hiddenState, long hiddenSize, optional MLGruCellOptions options = {});
};

dictionary MLInstanceNormalizationOptions {
  MLOperand scale;
  MLOperand bias;
  float epsilon = 1e-5;
  MLInputOperandLayout layout = "nchw";
};

partial interface MLGraphBuilder {
  MLOperand instanceNormalization(MLOperand input, 
                                optional MLInstanceNormalizationOptions options = {});
};

dictionary MLLeakyReluOptions {
  float alpha = 0.01;
};

partial interface MLGraphBuilder {
  MLOperand leakyRelu(MLOperand x, optional MLLeakyReluOptions options = {});
};

partial interface MLGraphBuilder {
  MLOperand matmul(MLOperand a, MLOperand b);
};

enum MLPaddingMode {
  "constant",
  "edge",
  "reflection",
  "symmetric"
};

dictionary MLPadOptions {
  MLPaddingMode mode = "constant";
  float value = 0;
};

partial interface MLGraphBuilder {
  MLOperand pad(MLOperand input, MLOperand padding, optional MLPadOptions options = {});
};

dictionary MLPool2dOptions {
  sequence<long> windowDimensions;
  sequence<long> padding;
  sequence<long> strides;
  sequence<long> dilations;
  MLAutoPad autoPad = "explicit";
  MLInputOperandLayout layout = "nchw";
};

partial interface MLGraphBuilder {
  MLOperand averagePool2d(MLOperand input, optional MLPool2dOptions options = {});
  MLOperand l2Pool2d(MLOperand input, optional MLPool2dOptions options = {});
  MLOperand maxPool2d(MLOperand input, optional MLPool2dOptions options = {});
};

dictionary MLReduceOptions {
  sequence<long> axes = null;
  boolean keepDimensions = false;
};

partial interface MLGraphBuilder {
  MLOperand reduceL1(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceL2(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceLogSum(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceLogSumExp(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceMax(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceMean(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceMin(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceProduct(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceSum(MLOperand input, optional MLReduceOptions options = {});
  MLOperand reduceSumSquare(MLOperand input, optional MLReduceOptions options = {});
};

enum MLInterpolationMode {
  "nearest-neighbor",
  "linear"
};

dictionary MLResampleOptions {
  MLInterpolationMode mode = "nearest-neighbor";
  sequence<float> scales;
  sequence<long> sizes;
};

partial interface MLGraphBuilder {
  MLOperand resample(MLOperand input, optional MLResampleOptions options = {});
};

partial interface MLGraphBuilder {
  MLOperand reshape(MLOperand input, sequence<long> newShape);
};

dictionary MLSliceOptions {
  sequence<long> axes;
};

partial interface MLGraphBuilder {
  MLOperand slice(MLOperand input, sequence<long> starts, sequence<long> sizes,
                optional MLSliceOptions options = {});
};

partial interface MLGraphBuilder {
  MLOperand softmax(MLOperand x);
};

dictionary MLSplitOptions {
  long axis = 0;
};

partial interface MLGraphBuilder {
  sequence<MLOperand> split(MLOperand input,
                          (unsigned long or sequence<unsigned long>) splits,
                          optional MLSplitOptions options = {});
};

dictionary MLSqueezeOptions {
  sequence<long> axes;
};

partial interface MLGraphBuilder {
  MLOperand squeeze(MLOperand input, optional MLSqueezeOptions options = {});
};

dictionary MLTransposeOptions {
  sequence<long> permutation;
};

partial interface MLGraphBuilder {
  MLOperand transpose(MLOperand input, optional MLTransposeOptions options = {});
};

dictionary MLInput {
  required (MLBufferView or WebGLTexture or GPUTexture) data;
  sequence<long> dimensions;
};

dictionary MLOutput {
  (MLBufferView or WebGLTexture or GPUTexture) data;
  sequence<long> dimensions;
};

typedef record<DOMString, MLInput> MLNamedInputs;
typedef record<DOMString, MLOutput> MLNamedOutputs;

[SecureContext, Exposed=(Window, DedicatedWorker)]
interface MLGraph {
  Promise<MLNamedOutputs> compute(MLNamedInputs inputs, 
                                  optional MLNamedOutputs outputs = {});
};

Issues Index

Describe the algorithm steps for this.[[context]] created from WebGLRenderingContext and GPUDevice.
clarify the usage of ArrayBufferView for float16. <https://github.com/webmachinelearning/webnn/issues/127>