sum

paddle.sum ( x, axis\=None, dtype\=None, keepdim\=False, name\=None ) [源代码]

该OP是对指定维度上的Tensor元素进行求和运算,并输出相应的计算结果。

参数:

  • x (Tensor)- 输入变量为多维Tensor,支持数据类型为float32,float64,int32,int64。

  • axis (int | list | tuple ,可选)- 求和运算的维度。如果为None,则计算所有元素的和并返回包含单个元素的Tensor变量,否则必须在

    sum - 图1

    范围内。如果

    sum - 图2

    ,则维度将变为 rank+axis[i]rank+axis[i] ,默认值为None。

  • dtype (str , 可选)- 输出变量的数据类型。若参数为空,则输出变量的数据类型和输入变量相同,默认值为None。

  • keepdim (bool)- 是否在输出Tensor中保留减小的维度。如 keepdim 为true,否则结果张量的维度将比输入张量小,默认值为False。

  • name (str , 可选)- 具体用法请参见 Name ,一般无需设置,默认值为None。

返回:

Tensor,在指定维度上进行求和运算的Tensor,数据类型和输入数据类型一致。

代码示例

  1. import numpy as np
  2. import paddle
  3. # x is a Tensor variable with following elements:
  4. # [[0.2, 0.3, 0.5, 0.9]
  5. # [0.1, 0.2, 0.6, 0.7]]
  6. # Each example is followed by the corresponding output tensor.
  7. x_data = np.array([[0.2, 0.3, 0.5, 0.9],[0.1, 0.2, 0.6, 0.7]]).astype('float32')
  8. x = paddle.to_tensor(x_data)
  9. out1 = paddle.sum(x) # [3.5]
  10. out2 = paddle.sum(x, axis=0) # [0.3, 0.5, 1.1, 1.6]
  11. out3 = paddle.sum(x, axis=-1) # [1.9, 1.6]
  12. out4 = paddle.sum(x, axis=1, keepdim=True) # [[1.9], [1.6]]
  13. # y is a Tensor variable with shape [2, 2, 2] and elements as below:
  14. # [[[1, 2], [3, 4]],
  15. # [[5, 6], [7, 8]]]
  16. # Each example is followed by the corresponding output tensor.
  17. y_data = np.array([[[1, 2], [3, 4]], [[5, 6], [7, 8]]]).astype('float32')
  18. y = paddle.to_tensor(y_data)
  19. out5 = paddle.sum(y, axis=[1, 2]) # [10, 26]
  20. out6 = paddle.sum(y, axis=[0, 1]) # [16, 20]