Sequence to sequence transformer model for expanding factors to polynomials in single variable. Training loss: Accuracy acheived ~ 99%