Use output of textcat.batch-train as input of textcat.teach

We are initiating a labeling effort and were planning to periodically batch-train on our labeled data and create a custom recipe that would load the most current version of this model for textcat.teach.

Are there downsides to such an approach?

No, in fact, this is a workflow we often recommend and a good way to continuously keep your models up to date :slightly_smiling_face:

1 Like