Package org.pytorch

Class Tensor

java.lang.Object
org.pytorch.Tensor
public abstract class Tensor
extends java.lang.Object
Representation of a Tensor. Behavior is similar to PyTorch's tensor objects.

Most tensors will be constructed as Tensor.fromBlob(data, shape), where data can be an array or a direct Buffer (of the proper subclass). Helper methods are provided to allocate buffers properly.

To access Tensor data, see dtype(), shape(), and various getDataAs* methods.

When constructing Tensor objects with data as an array, it is not specified whether this data is is copied or retained as a reference so it is recommended not to modify it after constructing. data passed as a Buffer is not copied, so it can be modified between Module calls to avoid reallocation. Data retrieved from Tensor objects may be copied or may be a reference to the Tensor's internal data buffer. shape is always copied.

  • Method Summary

    Modifier and Type Method Description
    static java.nio.ByteBuffer allocateByteBuffer​(int numElements)
    Allocates a new direct ByteBuffer with native byte order with specified capacity that can be used in fromBlob(ByteBuffer, long[]), fromBlobUnsigned(ByteBuffer, long[]).
    static java.nio.DoubleBuffer allocateDoubleBuffer​(int numElements)
    Allocates a new direct DoubleBuffer with native byte order with specified capacity that can be used in fromBlob(DoubleBuffer, long[]).
    static java.nio.FloatBuffer allocateFloatBuffer​(int numElements)
    Allocates a new direct FloatBuffer with native byte order with specified capacity that can be used in fromBlob(FloatBuffer, long[]).
    static java.nio.IntBuffer allocateIntBuffer​(int numElements)
    Allocates a new direct IntBuffer with native byte order with specified capacity that can be used in fromBlob(IntBuffer, long[]).
    static java.nio.LongBuffer allocateLongBuffer​(int numElements)
    Allocates a new direct LongBuffer with native byte order with specified capacity that can be used in fromBlob(LongBuffer, long[]).
    abstract DType dtype()  
    static Tensor fromBlob​(byte[] data, long[] shape)
    Creates a new Tensor instance with dtype torch.int8 with specified shape and data as array of bytes.
    static Tensor fromBlob​(float[] data, long[] shape)
    Creates a new Tensor instance with dtype torch.float32 with specified shape and data as array of floats.
    static Tensor fromBlob​(int[] data, long[] shape)
    Creates a new Tensor instance with dtype torch.int32 with specified shape and data as array of ints.
    static Tensor fromBlob​(long[] shape, double[] data)
    Creates a new Tensor instance with dtype torch.float64 with specified shape and data as array of doubles.
    static Tensor fromBlob​(long[] data, long[] shape)
    Creates a new Tensor instance with dtype torch.int64 with specified shape and data as array of longs.
    static Tensor fromBlob​(java.nio.ByteBuffer data, long[] shape)
    Creates a new Tensor instance with dtype torch.int8 with specified shape and data.
    static Tensor fromBlob​(java.nio.DoubleBuffer data, long[] shape)
    Creates a new Tensor instance with dtype torch.float64 with specified shape and data.
    static Tensor fromBlob​(java.nio.FloatBuffer data, long[] shape)
    Creates a new Tensor instance with dtype torch.float32 with specified shape and data.
    static Tensor fromBlob​(java.nio.IntBuffer data, long[] shape)
    Creates a new Tensor instance with dtype torch.int32 with specified shape and data.
    static Tensor fromBlob​(java.nio.LongBuffer data, long[] shape)
    Creates a new Tensor instance with dtype torch.int64 with specified shape and data.
    static Tensor fromBlobUnsigned​(byte[] data, long[] shape)
    Creates a new Tensor instance with dtype torch.uint8 with specified shape and data as array of bytes.
    static Tensor fromBlobUnsigned​(java.nio.ByteBuffer data, long[] shape)
    Creates a new Tensor instance with dtype torch.uint8 with specified shape and data.
    byte[] getDataAsByteArray()  
    double[] getDataAsDoubleArray()  
    float[] getDataAsFloatArray()  
    int[] getDataAsIntArray()  
    long[] getDataAsLongArray()  
    byte[] getDataAsUnsignedByteArray()  
    long numel()
    Returns the number of elements in this tensor.
    static long numel​(long[] shape)
    Calculates the number of elements in a tensor with the specified shape.
    long[] shape()
    Returns the shape of this tensor.

    Methods inherited from class java.lang.Object

    clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

  • Method Details

    • allocateByteBuffer

      public static java.nio.ByteBuffer allocateByteBuffer​(int numElements)
      Allocates a new direct ByteBuffer with native byte order with specified capacity that can be used in fromBlob(ByteBuffer, long[]), fromBlobUnsigned(ByteBuffer, long[]).
      Parameters:
      numElements - capacity (number of elements) of result buffer.

    • allocateIntBuffer

      public static java.nio.IntBuffer allocateIntBuffer​(int numElements)
      Allocates a new direct IntBuffer with native byte order with specified capacity that can be used in fromBlob(IntBuffer, long[]).
      Parameters:
      numElements - capacity (number of elements) of result buffer.

    • allocateFloatBuffer

      public static java.nio.FloatBuffer allocateFloatBuffer​(int numElements)
      Allocates a new direct FloatBuffer with native byte order with specified capacity that can be used in fromBlob(FloatBuffer, long[]).
      Parameters:
      numElements - capacity (number of elements) of result buffer.

    • allocateLongBuffer

      public static java.nio.LongBuffer allocateLongBuffer​(int numElements)
      Allocates a new direct LongBuffer with native byte order with specified capacity that can be used in fromBlob(LongBuffer, long[]).
      Parameters:
      numElements - capacity (number of elements) of result buffer.

    • allocateDoubleBuffer

      public static java.nio.DoubleBuffer allocateDoubleBuffer​(int numElements)
      Allocates a new direct DoubleBuffer with native byte order with specified capacity that can be used in fromBlob(DoubleBuffer, long[]).
      Parameters:
      numElements - capacity (number of elements) of result buffer.

    • fromBlobUnsigned

      public static Tensor fromBlobUnsigned​(byte[] data, long[] shape)
      Creates a new Tensor instance with dtype torch.uint8 with specified shape and data as array of bytes.
      Parameters:
      data - Tensor elements
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(byte[] data, long[] shape)
      Creates a new Tensor instance with dtype torch.int8 with specified shape and data as array of bytes.
      Parameters:
      data - Tensor elements
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(int[] data, long[] shape)
      Creates a new Tensor instance with dtype torch.int32 with specified shape and data as array of ints.
      Parameters:
      data - Tensor elements
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(float[] data, long[] shape)
      Creates a new Tensor instance with dtype torch.float32 with specified shape and data as array of floats.
      Parameters:
      data - Tensor elements
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(long[] data, long[] shape)
      Creates a new Tensor instance with dtype torch.int64 with specified shape and data as array of longs.
      Parameters:
      data - Tensor elements
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(long[] shape, double[] data)
      Creates a new Tensor instance with dtype torch.float64 with specified shape and data as array of doubles.
      Parameters:
      shape - Tensor shape
      data - Tensor elements

    • fromBlobUnsigned

      public static Tensor fromBlobUnsigned​(java.nio.ByteBuffer data, long[] shape)
      Creates a new Tensor instance with dtype torch.uint8 with specified shape and data.
      Parameters:
      data - Direct buffer with native byte order that contains Tensor.numel(shape) elements. The buffer is used directly without copying, and changes to its content will change the tensor.
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(java.nio.ByteBuffer data, long[] shape)
      Creates a new Tensor instance with dtype torch.int8 with specified shape and data.
      Parameters:
      data - Direct buffer with native byte order that contains Tensor.numel(shape) elements. The buffer is used directly without copying, and changes to its content will change the tensor.
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(java.nio.IntBuffer data, long[] shape)
      Creates a new Tensor instance with dtype torch.int32 with specified shape and data.
      Parameters:
      data - Direct buffer with native byte order that contains Tensor.numel(shape) elements. The buffer is used directly without copying, and changes to its content will change the tensor.
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(java.nio.FloatBuffer data, long[] shape)
      Creates a new Tensor instance with dtype torch.float32 with specified shape and data.
      Parameters:
      data - Direct buffer with native byte order that contains Tensor.numel(shape) elements. The buffer is used directly without copying, and changes to its content will change the tensor.
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(java.nio.LongBuffer data, long[] shape)
      Creates a new Tensor instance with dtype torch.int64 with specified shape and data.
      Parameters:
      data - Direct buffer with native byte order that contains Tensor.numel(shape) elements. The buffer is used directly without copying, and changes to its content will change the tensor.
      shape - Tensor shape

    • fromBlob

      public static Tensor fromBlob​(java.nio.DoubleBuffer data, long[] shape)
      Creates a new Tensor instance with dtype torch.float64 with specified shape and data.
      Parameters:
      data - Direct buffer with native byte order that contains Tensor.numel(shape) elements. The buffer is used directly without copying, and changes to its content will change the tensor.
      shape - Tensor shape

    • numel

      public long numel()
      Returns the number of elements in this tensor.

    • numel

      public static long numel​(long[] shape)
      Calculates the number of elements in a tensor with the specified shape.

    • shape

      public long[] shape()
      Returns the shape of this tensor. (The array is a fresh copy.)

    • dtype

      public abstract DType dtype()
      Returns:
      data type of this tensor.

    • getDataAsByteArray

      public byte[] getDataAsByteArray()
      Returns:
      a Java byte array that contains the tensor data. This may be a copy or reference.
      Throws:
      java.lang.IllegalStateException - if it is called for a non-int8 tensor.

    • getDataAsUnsignedByteArray

      public byte[] getDataAsUnsignedByteArray()
      Returns:
      a Java byte array that contains the tensor data. This may be a copy or reference.
      Throws:
      java.lang.IllegalStateException - if it is called for a non-uint8 tensor.

    • getDataAsIntArray

      public int[] getDataAsIntArray()
      Returns:
      a Java int array that contains the tensor data. This may be a copy or reference.
      Throws:
      java.lang.IllegalStateException - if it is called for a non-int32 tensor.

    • getDataAsFloatArray

      public float[] getDataAsFloatArray()
      Returns:
      a Java float array that contains the tensor data. This may be a copy or reference.
      Throws:
      java.lang.IllegalStateException - if it is called for a non-float32 tensor.

    • getDataAsLongArray

      public long[] getDataAsLongArray()
      Returns:
      a Java long array that contains the tensor data. This may be a copy or reference.
      Throws:
      java.lang.IllegalStateException - if it is called for a non-int64 tensor.

    • getDataAsDoubleArray

      public double[] getDataAsDoubleArray()
      Returns:
      a Java double array that contains the tensor data. This may be a copy or reference.
      Throws:
      java.lang.IllegalStateException - if it is called for a non-float64 tensor.