You can also indirectly use Datasets API in Python by firstly building Scala program with Datasets, creating JAR and running it Python.

Here is a tutorial on using Datasets in PySpark via Scala interoperability.