PySpark/Petastorm: How to generate and incorporate nn.Embedding() item?

Question

I am wondering if anyone has insight into how to access or generate new data when using Uber\'s petastorm?

What I Have Done:

Used vectorAs