InMemoryDataset new input parameter

[anthonysirico]

Can someone help define the init class shown here. I want to add a parameter called split, where its a float value like 0.25 that would be used later in the process section of the data.

class ImbCircuitDataset(InMemoryDataset):
    def __init__(self, root, transform=None, pre_transform=None,
                 pre_filter=None):
        super().__init__(root, transform, pre_transform, pre_filter)
        self.data, self.slices = torch.load(self.processed_paths[0])

so when I run dataset = ImbCircuitDataset('test'), i want to run dataset = ImbCircuitDataset(split=0.25, 'test') so i can automate my splitting process.

[EdisonLeeeee]

Did you mean that?

class ImbCircuitDataset(InMemoryDataset):
    def __init__(self, root, split, transform=None, pre_transform=None,
                 pre_filter=None):
        super().__init__(root, transform, pre_transform, pre_filter)
        self.split = split
        self.data, self.slices = torch.load(self.processed_paths[0])

[anthonysirico]

The returns AttributeError: 'ImbCircuitDataset' object has no attribute 'split'

[EdisonLeeeee]

Oh my bad. self.split should be defined before super().__init__, similar to https://github.com/pyg-team/pytorch_geometric/blob/master/torch_geometric/datasets/planetoid.py

[anthonysirico]

Thanks, friend. That worked!