Class TensorFloat

Represents data in a multidimensional array-like structure of floats.

Inheritance
Object
Tensor
TensorFloat
Inherited Members
Tensor.m_TensorOnDevice
Tensor.m_TensorAllocator
Tensor.m_Shape
Tensor.m_CacheIsDirty
Tensor.m_Disposed
Namespace: Unity.Sentis
Syntax
public class TensorFloat : Tensor, IDisposable

Constructors

TensorFloat(Single)

Initializes and returns a scalar tensor with the value of srcData.

Declaration
public TensorFloat(float srcData)
Parameters
Type Name Description
Single srcData

TensorFloat(TensorShape, Single[])

Initializes and returns a tensor with the specified shape and a float[] array of srcData data.

Declaration
public TensorFloat(TensorShape shape, float[] srcData)
Parameters
Type Name Description
TensorShape shape
Single[] srcData

TensorFloat(TensorShape, Single[], Int32)

Initializes and returns a tensor with specified shape and a float[] array of srcData data. Sentis reads srcData from dataStartIndex. You can provide an optional debug name.

srcData.Length - dataStartIndex must be bigger than or equal to shape.length.

Declaration
public TensorFloat(TensorShape shape, float[] srcData, int dataStartIndex = 0)
Parameters
Type Name Description
TensorShape shape
Single[] srcData
Int32 dataStartIndex

Properties

dataType

The data type of the elements of the tensor.

Declaration
public override DataType dataType { get; }
Property Value
Type Description
DataType
Overrides

isCacheNull

Declaration
protected override bool isCacheNull { get; }
Property Value
Type Description
Boolean
Overrides

Item[Int32]

Returns the tensor element at offset d0.

If the tensor is the result of computation on a different device, the method creates a blocking read.

Declaration
public float this[int d0] { get; set; }
Parameters
Type Name Description
Int32 d0
Property Value
Type Description
Single

Item[Int32, Int32]

Returns the tensor element at offset (d1, d0), which is position d1 * stride0 + d0.

If the tensor is the result of computation on a different device, the method creates a blocking read.

Declaration
public float this[int d1, int d0] { get; set; }
Parameters
Type Name Description
Int32 d1
Int32 d0
Property Value
Type Description
Single

Item[Int32, Int32, Int32]

Returns the tensor element at offset (d2, d1, d0), which is position d2 * stride1 + d1 * stride0 + d0.

If the tensor is the result of computation on a different device, the method creates a blocking read.

Declaration
public float this[int d2, int d1, int d0] { get; set; }
Parameters
Type Name Description
Int32 d2
Int32 d1
Int32 d0
Property Value
Type Description
Single

Item[Int32, Int32, Int32, Int32]

Returns the tensor element at offset (d3, d2, d1, d0), which is position d3 * stride2 + d2 * stride1 + d1 * stride0 + d0 in this tensor.

If the tensor is the result of computation on a different device, the method creates a blocking read.

Declaration
public float this[int d3, int d2, int d1, int d0] { get; set; }
Parameters
Type Name Description
Int32 d3
Int32 d2
Int32 d1
Int32 d0
Property Value
Type Description
Single

Item[Int32, Int32, Int32, Int32, Int32]

Returns the tensor element at offset (d4, d3, d2, d1, d0), which is position d4 * stride3 + d3 * stride2 + d2 * stride1 + d1 * stride0 + d0.

If the tensor is the result of computation on a different device, the method creates a blocking read.

Declaration
public float this[int d4, int d3, int d2, int d1, int d0] { get; set; }
Parameters
Type Name Description
Int32 d4
Int32 d3
Int32 d2
Int32 d1
Int32 d0
Property Value
Type Description
Single

Item[Int32, Int32, Int32, Int32, Int32, Int32]

Returns the tensor element at offset (d5, d4, d3, d2, d1, d0), which is position d5 * stride4 + d4 * stride3 + d3 * stride2 + d2 * stride1 + d1 * stride0 + d0.

If the tensor is the result of computation on a different device, the method creates a blocking read.

Declaration
public float this[int d5, int d4, int d3, int d2, int d1, int d0] { get; set; }
Parameters
Type Name Description
Int32 d5
Int32 d4
Int32 d3
Int32 d2
Int32 d1
Int32 d0
Property Value
Type Description
Single

Item[Int32, Int32, Int32, Int32, Int32, Int32, Int32]

Returns the tensor element at offset (d6, d5, d4, d3, d2, d1, d0), which is position d6 * stride5 + d5 * stride4 + d4 * stride3 + d3 * stride2 + d2 * stride1 + d1 * stride0 + d0.

If the tensor is the result of computation on a different device, the method creates a blocking read.

Declaration
public float this[int d6, int d5, int d4, int d3, int d2, int d1, int d0] { get; set; }
Parameters
Type Name Description
Int32 d6
Int32 d5
Int32 d4
Int32 d3
Int32 d2
Int32 d1
Int32 d0
Property Value
Type Description
Single

Item[Int32, Int32, Int32, Int32, Int32, Int32, Int32, Int32]

Returns the tensor element at offset (d7, d6, d5, d4, d3, d2, d1, d0), which is position d7 * stride6 + d6 * stride5 + d5 * stride4 + d4 * stride3 + d3 * stride2 + d2 * stride1 + d1 * stride0 + d0.

If the tensor is the result of computation on a different device, the method creates a blocking read.

Declaration
public float this[int d7, int d6, int d5, int d4, int d3, int d2, int d1, int d0] { get; set; }
Parameters
Type Name Description
Int32 d7
Int32 d6
Int32 d5
Int32 d4
Int32 d3
Int32 d2
Int32 d1
Int32 d0
Property Value
Type Description
Single

Methods

ClearCache()

Declaration
protected override void ClearCache()
Overrides

DeepCopy()

Returns a deep copy of the current Tensor.

Declaration
public override Tensor DeepCopy()
Returns
Type Description
Tensor
Overrides

PrepareCacheForAccess(Boolean)

Read data from the device and write it to the cache.

The default value of blocking is true, which means this method is a blocking read.

When the value of blocking is false, the read is non-blocking. You can keep calling the method to get the status of the asynchronous download. You can access the tensor data when the method returns true.

Declaration
public override bool PrepareCacheForAccess(bool blocking = true)
Parameters
Type Name Description
Boolean blocking
Returns
Type Description
Boolean
Overrides

ShallowReshape(TensorShape)

Returns a shallow copy of the Tensor with a new shape. The copy shares data storage with the original tensor.

newShape.length must be equal to this.shape.length.

Declaration
public override Tensor ShallowReshape(TensorShape newShape)
Parameters
Type Name Description
TensorShape newShape
Returns
Type Description
Tensor
Overrides

ToReadOnlyArray()

Returns the cached linear memory representation of the data in this tensor.

If the tensor is the result of computation on a different device, the method creates a blocking read.

If you modify the contents of the returned array, the behaviour is undefined.

Declaration
public float[] ToReadOnlyArray()
Returns
Type Description
Single[]

UploadIfDirty()

Declaration
protected override void UploadIfDirty()
Overrides

Zeros(TensorShape)

Initializes and returns a tensor with the specified shape and filled with 0.

Declaration
public static TensorFloat Zeros(TensorShape shape)
Parameters
Type Name Description
TensorShape shape
Returns
Type Description
TensorFloat

Extension Methods

Did you find this page useful? Please give it a rating: