Is there a more efficient way to accomplish the following? (for general X, sizes)
sizes = [3, 7, 5, 9]
X = torch.ones(sum(sizes))
Y = torch.tensor([s.sum() for s in torch.split(X, sizes)])
For instance with numpy I can do the following
indices = np.cumsum(+sizes)[:-1]
Y = np.add.reduceat(X, indices.tolist())
Is there a pytorch equivalent for this?