TypeError: torch.index_select received an invalid combination of arguments - got (torch.FloatTensor, int, torch.cuda.LongTensor), but expected (torch.FloatTensor source, int dim, torch.LongTensor index)

[anhquan0412]

TypeError: torch.index_select received an invalid combination of arguments - got (torch.FloatTensor, int, torch.cuda.LongTensor), but expected (torch.FloatTensor source, int dim, torch.LongTensor index)

[anhquan0412]

Hi,
The error is because some part of my notebook is not run on a GPU machine so I removed the 'cuda' function. Since you run it on GPU you should add .cuda() to the model to put the model on GPU, i.e in this case it will be

em_model = EmbeddingMult(n_factors,cf.n_users,cf.n_items).cuda()
model = CollabFilterModel(em_model)
learner = CollabFilterLearner(cf.get_data(val_idxs,bs),model,opt_fn=optim.Adam)
learner.fit(1e-2, 3, wds=1e-4, cycle_len=1,cycle_mult=2)

Let me know if it helps!

[eoc1991]

It works, thank you very much. :)

[anhquan0412]

No problem. Glad to help!