Dynamic Dataframe possible?

data
1

I’m not entirely certain how to accomplish what I want to accomplish with pytorch yet, but I was wondering if its possible to have a base dataframe that is for example: unlimited rows, and 3 columns

column 1 has 1 value, column 2 is an array that can hold any amount of arrays, and column 3 is an array that can hold any amount of arrays

df = pd.DataFrame([
    ['Col 1', 'Col 2', 'Col 3'],
    [
        1,
        [
            [
                1,
                [
                    [1, 1, 1, 1],
                    [1, 1, 1, 1]
                ]
            ],
            [
                2,
                [
                    [2, 2, 2, 2],
                    [2, 2, 2, 2]
                ]
            ]
        ],
        [
            [
                1,
                [1, 1]
            ],
            [
                2,
                [2, 2]
            ]
        ]
    ]
])