Changing the Y Value of the baseline data (time crunch on this one)

[anthonysirico]

If i have an InMemoryDataset, how can I change or add each graph's Y value? There is a reason behind this, just please humor me.

dataset = MyDataset(path) # currently contains no Y values for a reason

performance = dataset.data.performance.tolist()

median_performance = np.median(performance)

for i in range(len(dataset)):
    if dataset[i].performance < median_performance:
        dataset[i].Y = torch.tensor([1])
    else:
        dataset[i].Y = torch.tensor([0])

I am doing this in a loop because performance changes every time. But, when it gets to the training, I get:

AttributeError: 'GlobalStorage' object has no attribute 'Y'

[rusty1s]

The recommended way to do this is through the usage of transforms, e.g.:

class MyTransform:
     def __init__(self, median_performance):
         self.median_performance = median_performance
     
     def __call__(self, data):
          if data.performance < median_performance:
               data.y = 1
          else:
               data.y = 0
          return data
          
dataset.transform = MyTransform(median_performance)
     

[anthonysirico]

Thank you!