Classification vs (Embed + SVM)

Is it better to classify using the Embed + SVM approach (as documented in the Cohere Docs) or to use Classify model directly?

In other words, is the classify model using advantage of few shot learning examples in any way?

Hey! We’ll be releasing an out of the box Embed + SVM approach (identical experience to the current Classify except you just upload data) in a few weeks. The current version of Classify certainly makes use of the learning examples and it’s a great way to get started. But for the best possible performance we recommend the documented Embed + SVM approach.