Extract tensors from custom class into one big tensor

1

load_chunks
load_chunks1050×264 85 KB

I need to extract every 0th and 2nd Tensor from chunks.fulldata and put them in two separate Tensors (features and targets) that can be loaded into Pytorch’s DataLoader - what’s the best way to do this? I’m sure there’s some simple way to do this, but I’m new to this, so I’m kind of lost here

2

Based on your screenshot it looks like fulldata is a tuple containing 43 tensors.
I’m not sure, how you would like to split this tuple, but this would be one way:

fulldata = tuple(torch.randn(1) for _ in range(100))
features = torch.stack(fulldata[::2])
targets = torch.stack(fulldata[1::2])
1 Like