When I try to run the example notebook svi_compare, the cell svi.fit(train_data, val_data=None) causes an error which is traceable to the repeat_rows call from the stochiastic_elbo function in vae.py. The problem seems to be that the DataLoader elements are lists where each list item contains two tensors, but repeat_rows expects just a tensor. However, other parts of the code do expect the DataLoader to return lists. Is there a fix for this?
When I try to run the example notebook svi_compare, the cell svi.fit(train_data, val_data=None) causes an error which is traceable to the repeat_rows call from the stochiastic_elbo function in vae.py. The problem seems to be that the DataLoader elements are lists where each list item contains two tensors, but repeat_rows expects just a tensor. However, other parts of the code do expect the DataLoader to return lists. Is there a fix for this?