Sequence to sequence transformer model for expanding factors to polynomials in single variable.

Training loss:

Accuracy acheived ~ 99%