I am looking for a generic way to construct a block tensor from a specified arrangement of multiple tensors. For example, constructing the tensor `[[A, B], [C, D]] `

where `A`

, `B`

, `C`

, `D`

are all tensors. In `numpy`

simply `np.block( [[A, B], [C, D]])`

does the trick. Is there any similar function in `pytorch`

to achieve the same?

Doc for `numpy.block`

: numpy.block - manual