在Python中计算张量的点积

给定两个张量a和b，以及包含两个array_like对象(a_axes, b_axes)的数组对象，在由a_axes和b_axes指定的轴上求a和b的元素(分量)的乘积之和。第三个参数可以是单个非负整数类型的标量N；如果是这样，那么a的最后N个维度和b的前N个维度将进行求和。

要计算张量的点积，在Python中使用numpy.tensordot()方法。参数a和b是要进行“点积”的张量。参数axes是整数类型，如果是整数N，则按顺序将a的最后N个轴和b的前N个轴进行求和。相应轴的大小必须匹配。

步骤

首先，导入所需的库−

import numpy as np

创建两个numpy的3D数组，使用array()方法。

arr1 = np.arange(60.).reshape(3,4,5)
arr2 = np.arange(24.).reshape(4,3,2)

显示数组 –

print("Array1...\n",arr1)
print("\nArray2...\n",arr2)

检查两个数组的维度 –

print("\nDimensions of Array1...\n",arr1.ndim)
print("\nDimensions of Array2...\n",arr2.ndim)

检查两个数组的形状 –

print("\nShape of Array1...\n",arr1.shape)
print("\nShape of Array2...\n",arr2.shape)

要计算张量的点积，可以使用Python中的numpy.tensordot()方法。参数a、b是要“dot”的张量 −

print("\nTensor dot product...\n", np.tensordot(arr1,arr2, axes=([1,0],[0,1])))

示例

import numpy as np

# Creating two numpy 3D arrays using the array() method
arr1 = np.arange(60.).reshape(3,4,5)
arr2 = np.arange(24.).reshape(4,3,2)

# Display the arrays
print("Array1...\n",arr1)
print("\nArray2...\n",arr2)

# Check the Dimensions of both the arrays
print("\nDimensions of Array1...\n",arr1.ndim)
print("\nDimensions of Array2...\n",arr2.ndim)

# Check the Shape of both the arrays
print("\nShape of Array1...\n",arr1.shape)
print("\nShape of Array2...\n",arr2.shape)

# To compute the tensor dot product, use the numpy.tensordot() method in Python
# The a, b parameters are Tensors to “dot”.
print("\nTensor dot product...\n", np.tensordot(arr1,arr2, axes=([1,0],[0,1])))

输出

Array1...
[[[ 0. 1. 2. 3. 4.]
[ 5. 6. 7. 8. 9.]
[10. 11. 12. 13. 14.]
[15. 16. 17. 18. 19.]]

[[20. 21. 22. 23. 24.]
[25. 26. 27. 28. 29.]
[30. 31. 32. 33. 34.]
[35. 36. 37. 38. 39.]]

[[40. 41. 42. 43. 44.]
[45. 46. 47. 48. 49.]
[50. 51. 52. 53. 54.]
[55. 56. 57. 58. 59.]]]

Array2...
[[[ 0. 1.]
[ 2. 3.]
[ 4. 5.]]

[[ 6. 7.]
[ 8. 9.]
[10. 11.]]

[[12. 13.]
[14. 15.]
[16. 17.]]

[[18. 19.]
[20. 21.]
[22. 23.]]]

Dimensions of Array1...
3

Dimensions of Array2...
3

Shape of Array1...
(3, 4, 5)

Shape of Array2...
(4, 3, 2)

Tensor dot product...
[[4400. 4730.]
[4532. 4874.]
[4664. 5018.]
[4796. 5162.]
[4928. 5306.]]