You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, In a research project, I'm trying to give one token to GPT2 and then output the exact ranking you get in the layer_predictions function. Eventually, I intend to combine this output with a translation model output (both of the outputs should have the length of the vocabulary) and then get one token chosen by linearly combining both models ranking (this method is called shallow fusion in NMT).
Is it possible that I use your package to do this? I'd really appreciate your help and advice on how this can be possible!
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hi @jalammar,
Currently, In a research project, I'm trying to give one token to GPT2 and then output the exact ranking you get in the layer_predictions function. Eventually, I intend to combine this output with a translation model output (both of the outputs should have the length of the vocabulary) and then get one token chosen by linearly combining both models ranking (this method is called shallow fusion in NMT).
Is it possible that I use your package to do this? I'd really appreciate your help and advice on how this can be possible!
Beta Was this translation helpful? Give feedback.
All reactions