RuntimeError: a leaf Variable that requires grad is being used in an in-place operation

[abedkurdi]

Hello there, I have managed to make the package installation work, and while I am trying to reproduce the results in your paper using the following commands:

from gears import PertData, GEARS

# get data pert_data = PertData('./data') # load dataset in paper: norman, adamson, dixit. pert_data.load(data_name = 'norman') # specify data split pert_data.prepare_split(split = 'simulation', seed = 1) # get data loader with batch size pert_data.get_dataloader(batch_size = 32, test_batch_size = 128)

# set up and train a model gears_model = GEARS(pert_data, device = 'cpu') gears_model.model_initialize(hidden_size = 64) gears_model.train(epochs = 20)

When I reach the last command gears_model.train(epochs = 20), the program is throwing the following error: RuntimeError: a leaf Variable that requires grad is being used in an in-place operation. Do you have any idea how to overcome this problem?

Note. I am using cpu as device, I don't have CUDA.

Thanks in advance.

[yhr91]

Thanks for your comment and sorry for the late reply. I believe this error only occurs when the model is run using a CPU. In this case, switching out

losses += torch.sum((pred_p - y_p)**(2 + gamma))/pred_p.shape[0]/pred_p.shape[1]

with

losses = losses + torch.sum((pred_p - y_p)**(2 + gamma))/pred_p.shape[0]/pred_p.shape[1]

seems to solve the problem. Let me know if that doesn't work for you

https://github.com/snap-stanford/GEARS/blob/0a7d43b6d0cf17b03834495c08f0c450de1c0220/gears/utils.py#L296