Defining an embedding is quite similar to defining a feature. It’s associated with an entity, and the data is stored in a data set. An embedding must be a vector of floats, so all that needs specification is the dimensions of the embedding and the vector DB in which it’s stored. Here’s an example: 
comment_embeddings = ff.Embedding(
    vectorize_comments[["PK", "Vector"]],
    dims=384,
    vector_db=pinecone,
    description="Embeddings created from speakers' comments in episodes",
    variant="v1"
)
Additionally, we provide a more detailed end-to-end example of building with embeddings.