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
To my knowledge, the target sentence is generated word by word. Usually, the next word is calculated based on the input sequence and the decoder state.
I'd like to pass all previously generated target words to the decoder as well (I'll use a RNN to iterate over them, so the varying sequence length isn't a problem).
So if I'm translating the german "Ich bin hungrig", and already generated "I am", the decoder should calculate decoder(previous_decoder_state, input_sequence, "I am") = hungry
But how do I access the previously generated target words?
I'd like to use the already translated part of the target for the calculation of the following part.
Is this supported by blocks?
The text was updated successfully, but these errors were encountered: