torch.vstack — PyTorch 2.0 documentation (original) (raw)

torch.vstack(tensors, *, out=None) → Tensor¶

Stack tensors in sequence vertically (row wise).

This is equivalent to concatenation along the first axis after all 1-D tensors have been reshaped by torch.atleast_2d().

Parameters:

tensors (sequence of Tensors) – sequence of tensors to concatenate

Keyword Arguments:

out (Tensor, optional) – the output tensor.

Example:

a = torch.tensor([1, 2, 3]) b = torch.tensor([4, 5, 6]) torch.vstack((a,b)) tensor([[1, 2, 3], [4, 5, 6]]) a = torch.tensor([[1],[2],[3]]) b = torch.tensor([[4],[5],[6]]) torch.vstack((a,b)) tensor([[1], [2], [3], [4], [5], [6]])