Random split alternative

1

I do not want to split data into train, test and validation randomly but split them sequentially. What can I do?

Please help

2

I’m not sure I understand the issue correctly, but a “sequential” split could just be done via indexing/slicing:

data = torch.arange(100)
train, val, test = data[:80], data[80:90], data[90:]
1 Like