# 比较与torch.cumprod的功能差异

## torch.cumprod ```text torch.cumprod(input, dim, *, dtype=None, out=None) -> Tensor ``` 更多内容详见[torch.cumprod](https://pytorch.org/docs/1.8.1/generated/torch.cumprod.html)。 ## mindspore.ops.cumprod ```text mindspore.ops.cumprod(input, dim, dtype=None) -> Tensor ``` 更多内容详见[mindspore.ops.cumprod](https://www.mindspore.cn/docs/zh-CN/r2.0.0-alpha/api_python/ops/mindspore.ops.cumprod.html)。 ## 差异对比 PyTorch：计算`input`沿着指定维度`dim`的元素累计积,`dtype`参数用于转换输入的数据格式。 MindSpore：MindSpore此API实现功能与PyTorch基本一致。 | 分类 | 子类 |PyTorch | MindSpore | 差异 | | --- | --- | --- | --- |---| | 参数| 参数1 | input | input | - | | | 参数2 | dim | dim | - | | | 参数3 | dtype | dtype | - | | | 参数4 | out | - | 不涉及 | ### 代码示例1 > PyTorch和MindSpore中此API实现相同功能。 ```python # PyTorch import torch x = torch.tensor([[1, 2, 3], [4, 5, 6], [7, 8, 9]],dtype=int) out = torch.cumprod(x, 0) out = out.detach().numpy() print(out) # [[ 1 2 3] # [ 4 10 18] # [ 28 80 162]] # MindSpore import mindspore from mindspore import Tensor import mindspore.ops as ops x = Tensor([[1, 2, 3], [4, 5, 6], [7, 8, 9]]).astype('int32') out = ops.cumprod(x, 0) print(out) # [[ 1 2 3] # [ 4 10 18] # [ 28 80 162]] ``` ### 代码示例2 > PyTorch和MindSpore中此API参数`dtype`用于转换`input`的数据类型。 ```python # PyTorch import torch x = torch.tensor([[1, 2, 3], [4, 5, 6], [7, 8, 9]],dtype=int) out = torch.cumprod(x, 0, dtype=float) out = out.detach().numpy() print(out) # [[ 1. 2. 3.] # [ 4. 10. 18.] # [ 28. 80. 162.]] # MindSpore import mindspore from mindspore import Tensor import mindspore.ops as ops x = Tensor([[1, 2, 3], [4, 5, 6], [7, 8, 9]]).astype('int32') out = ops.cumprod(x, 0, dtype=mindspore.float32) print(out) # [[ 1. 2. 3.] # [ 4. 10. 18.] # [ 28. 80. 162.]] ```