Class Worker

Represents a worker that allows you to execute neural networks (models).

Worker abstracts implementation details on different hardware devices such as the CPU and the GPU. Worker lets you do the following:

• Specify inputs.
• Schedule the work.
• Get outputs.

Internally, Worker translates the neural network from a Model into a set of operations, then sends the operations to the hardware device for asynchronous execution.

Inheritance

object

Worker

Implements

IDisposable

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Namespace: Unity.Sentis

Assembly: Unity.Sentis.dll

Syntax

public class Worker : IDisposable

Constructors

Worker(Model, BackendType)

Creates and initializes a Worker with a model on a backend type.

Declaration

public Worker(Model model, BackendType backendType)

Parameters

Type	Name	Description
Model	model	The model for the worker.
BackendType	backendType	The backend type for the worker.

Worker(Model, DeviceType)

Creates and initializes a Worker with a model on a backend type.

The fastest backend type for the device type will be chosen.

Declaration

public Worker(Model model, DeviceType deviceType)

Parameters

Type	Name	Description
Model	model	The model for the worker.
DeviceType	deviceType	The device type for the worker.

Properties

backendType

Returns the backend type of the worker.

Declaration

public BackendType backendType { get; }

Property Value

Type	Description
BackendType

Methods

CopyOutput(int, ref Tensor)

Schedule a copy of the output tensor at an index into a tensor.

If, the input tensor is null, Sentis will allocate a new one If not, the input tensor dataType must match the output and have large enough capacity for the output shape.

Sentis reshapes the provided tensor if the shape does not match.

Declaration

public void CopyOutput(int index, ref Tensor tensor)

Parameters

Type	Name	Description
int	index	The index of the output.
Tensor	tensor	The tensor to copy the output into.

CopyOutput(string, ref Tensor)

Schedule a copy of the output tensor with a name into a tensor.

If, the input tensor is null, Sentis will allocate a new one If not, the input tensor dataType must match the output and have large enough capacity for the output shape.

Sentis reshapes the provided tensor if the shape does not match.

Declaration

public void CopyOutput(string name, ref Tensor tensor)

Parameters

Type	Name	Description
string	name	The name of the output.
Tensor	tensor	The tensor to copy the output into.

Dispose()

Disposes of the worker and any associated memory.

Declaration

public void Dispose()

~Worker()

Finalizes the Worker.

Declaration

protected ~Worker()

PeekOutput()

Returns a reference to the default output tensor. This is non-blocking.

For models with more than one output this returns a reference to the first output tensor.

The reference is valid only until you call Execute() or Dispose() on the worker.

Declaration

public Tensor PeekOutput()

Returns

Type	Description
Tensor	The output tensor reference.

PeekOutput(int)

Returns a reference an output tensor. This is non-blocking.

The reference is valid only until you call Execute() or Dispose() on the worker.

Declaration

public Tensor PeekOutput(int index)

Parameters

Type	Name	Description
int	index	The index of the output tensor to peek.

Returns

Type	Description
Tensor	The output tensor reference.

PeekOutput(string)

Returns a reference to an output tensor. This is non-blocking.

The reference is valid only until you call Execute() or Dispose() on the worker.

Declaration

public Tensor PeekOutput(string name)

Parameters

Type	Name	Description
string	name	The name of the output tensor to peek.

Returns

Type	Description
Tensor	The output tensor reference.

Schedule()

Schedules the execution of the model on the worker. This is non-blocking.

Declaration

public void Schedule()

Schedule(Tensor)

Schedules the execution of the model on the worker. This is non-blocking.

Declaration

public void Schedule(Tensor input)

Parameters

Type	Name	Description
Tensor	input	The tensor to set to the default input of the model.

Schedule(params Tensor[])

Schedules the execution of the model on the worker. This is non-blocking.

Declaration

public void Schedule(params Tensor[] inputs)

Parameters

Type	Name	Description
Tensor[]	inputs	The tensors to set.

ScheduleIterable()

Schedules the execution of the model in parts. This is non-blocking.

To schedule the execution of the next layer of the model, call MoveNext on the IEnumerator object this method returns.

Declaration

public IEnumerator ScheduleIterable()

Returns

Type	Description
IEnumerator	The IEnumerator for iterating the scheduling.

ScheduleIterable(Tensor)

Schedules the execution of the model in parts. This is non-blocking.

To schedule the execution of the next layer of the model, call MoveNext on the IEnumerator object this method returns.

Declaration

public IEnumerator ScheduleIterable(Tensor input)

Parameters

Type	Name	Description
Tensor	input	The tensor to set to the default input of the model.

Returns

Type	Description
IEnumerator	The IEnumerator for iterating the scheduling.

ScheduleIterable(params Tensor[])

Schedules the execution of the model in parts. This is non-blocking.

To schedule the execution of the next layer of the model, call MoveNext on the IEnumerator object this method returns.

Declaration

public IEnumerator ScheduleIterable(params Tensor[] inputs)

Parameters

Type	Name	Description
Tensor[]	inputs	The tensors to set.

Returns

Type	Description
IEnumerator	The IEnumerator for iterating the scheduling.

SetInput(int, Tensor)

Sets an input tensor with an index.

Declaration

public void SetInput(int index, Tensor input)

Parameters

Type	Name	Description
int	index	The index of the input to set.
Tensor	input	The input tensor.

SetInput(string, Tensor)

Sets an input tensor with a name.

Declaration

public void SetInput(string name, Tensor input)

Parameters

Type	Name	Description
string	name	The name of the input to set.
Tensor	input	The input tensor.

Implements

IDisposable