When the torch.masked_select operator passes in the same parameters, it behaves differently on CPU and GPU.

[triumph-wangyuyang]

🐛 Describe the bug

Test on the CPU:

import torch
input = torch.rand([], dtype=torch.float64)
out = torch.logcumsumexp(input, dim=1)
print(out)

result: IndexError: Dimension out of range (expected to be in range of [-1, 0], but got 1)

Test on the CPU:

import torch
input = torch.rand([], dtype=torch.float64).cuda()
out = torch.logcumsumexp(input, dim=1)
print(out)

result: IndexError: tensor(0.4230, device='cuda:0', dtype=torch.float64)

When the torch.masked_select operator passes in the same parameters, it behaves differently on CPU and GPU.

Versions

pytorch: 1.12.1
Python version: 3.8
CUDA/cuDNN version: cuDNN 11.3
GPU models and configuration: RTX3060
Operating System：Windows

[jingxu10]

It seems like the behavior on CPU is as expected.