How do I make a prediction when using FastAi collaborative filtering?

Question

Based on the model that Jeremy built in lesson 7 I am trying to predict the score of a specific user and film (user 460 and film fly away home 1996.

I tried the following approach

new_df = pd.DataFrame({'user':[460],'title':['Fly Away Home (1996)']})
dl = learn.dls.test_dl(new_df)
learn.get_preds(dl=dl)

But got this error message

IndexError: too many indices for tensor of dimension 0