Python `in` operator in torch

1

I am accidentally using in operator to check whether a two-dimensional torch contains a single vector. But the results are not the same as expected?

Intuitively, test 2 3 4 should also fail but they didn’t.

My platform: windows 10, torch1.12.0+cpu.

    import torch
    b = torch.LongTensor([
        [1, 2, 3],
        [4, 5, 6],
    ])
    # test 1 
    assert torch.LongTensor([1, 2, 3]) in b  # pass as expected 
    # test 2
    assert torch.LongTensor([1, 2, 4]) in b  # why pass ?
    # test 3
    assert torch.LongTensor([1, 2, 8]) in b  # why pass ?
    # test 4
    assert torch.LongTensor([2, 2, 6]) in b  # why pass ?
    # test 5 
    assert torch.LongTensor([8, 8, 8]) in b  # fail as expected