PyTorch torch.cumprod 函数
Pytorch torch 参考手册torch.cumprod 是 PyTorch 中用于计算累积积的函数。它返回沿指定维度的累积乘积，即从开始到当前位置所有元素的乘积。
函数定义
torch.cumprod(input, dim, dtype=None)
使用示例
实例
import torch
# 计算累积积
x = torch.tensor([1, 2, 3, 4, 5])
result = torch.cumprod(x, dim=0)
print("输入:", x)
print("累积积:", result)
# 输出: tensor([1, 2, 6, 24, 120])
# 说明: 1, 1*2=2, 1*2*3=6, 1*2*3*4=24, 1*2*3*4*5=120
# 2 维张量
x = torch.tensor([[1, 2, 3], [4, 5, 6]], dtype=torch.float32)
# 按列累积
result_col = torch.cumprod(x, dim=0)
print("n按列累积积:")
print(result_col)
# tensor([[ 1, 2, 3],
# [ 4, 10, 18]])
# 按行累积
result_row = torch.cumprod(x, dim=1)
print("n按行累积积:")
print(result_row)
# tensor([[ 1, 2, 6],
# [ 4, 20, 120]])
# 计算累积积
x = torch.tensor([1, 2, 3, 4, 5])
result = torch.cumprod(x, dim=0)
print("输入:", x)
print("累积积:", result)
# 输出: tensor([1, 2, 6, 24, 120])
# 说明: 1, 1*2=2, 1*2*3=6, 1*2*3*4=24, 1*2*3*4*5=120
# 2 维张量
x = torch.tensor([[1, 2, 3], [4, 5, 6]], dtype=torch.float32)
# 按列累积
result_col = torch.cumprod(x, dim=0)
print("n按列累积积:")
print(result_col)
# tensor([[ 1, 2, 3],
# [ 4, 10, 18]])
# 按行累积
result_row = torch.cumprod(x, dim=1)
print("n按行累积积:")
print(result_row)
# tensor([[ 1, 2, 6],
# [ 4, 20, 120]])