Spancat scripted training without CLI commands

[harshchavhan]

facing issues in passing annotated jsonl format training-dataset file which is generated by prodigy db-out command.
Trying to train prodigy annotate dataset directly on spacy without using CLI commands @polm

[ljvmiranda921]

Hi @harshchavhan ,

We appreciate not tagging maintainers directly in the discussion forum. If you already exported your data from Prodigy using db-out, then you can use the data-to-spacy command to convert the data into the spaCy format. From there, you can use spacy train to perform training. You can check the training documentation for more info.

facing issues in passing annotated jsonl format training-dataset file

Can you copy-paste the text of the error or issue?

It's a bit unclear what you meant by not using CLI commands. If you want to perform training programmatically, one way is to see how it's done in spacy train and adapt it into your codebase. Under the hood, it calls the train function for model training.