fix issue #664 - inverted token and pos emb layers

[casinca]

fixes #664

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[d-kleine]

Maybe it would make sense removing everything after encoded_text = tokenizer.encode(raw_text) as it's not relevant for exercise, what do you think @rasbt?

[rasbt]

Thanks for the fix @casinca . And I also agree with you @d-kleine , the lines seem redundant and I removed them.