1
$\begingroup$

Why is it called a Seq2Seq model if the output is just a number?

For example, if you are trying to predict a movie's recommendation, and you are inputting a sequence of users and their ratings, shouldn't it be a Seq2Number model since you're only predicting 1 rating at a time?

$\endgroup$

1 Answer 1

4
$\begingroup$

The premise of your question is wrong. A model that goes from a sequence to a single prediction is simply NOT called a sequence to sequence to model.

The model type you are describing is called a sequence encoder. An example would be sentiment prediction, where we input a sequence of text and output a number.

Similarly, a model that goes from a fixed size value to a sequence is called a sequence decoder. An example would be image captioning.

If a model inputs one sequence and outputs another, as in machine translation, it is called a sequence to sequence model, and consists of both an encoder and a decoder.

If you saw different terminology, it was either mislabeled or misunderstood.

$\endgroup$
1
  • 1
    $\begingroup$ Thank you for the clarification! I definitely misunderstood bc I couldn't find definitions as clear as this one. $\endgroup$
    – Katsu
    Feb 24, 2023 at 18:25

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .