textcat training with only one label

Hi I have used textcat to manually annotate my dataset if something should be categorised as self-service or not:
python -m prodigy textcat.manual sd_2023_selfs data_sd_2023.jsonl --label Self_Service
python -m prodigy db-out sd_2023_selfs > data_sd_2023_annot_selfs.jsonl
After few hundreds examples I wanted to train the model, so that at the end I could use textcat.corrrect to check how well was the training. As in the script below:
python -m prodigy train ./models_cat_selfs --textcat sd_2023_selfs
python -m prodigy textcat.correct sd_2023_selfs models_cat_selfs/model-best ./data/data_sd_2023_annot_selfs.jsonl --label Self_Service
But as I try to start train part I got the error that The 'textcat' component requires at least two labels because it uses mutually exclusive classes where exactly one label is True for each doc.

Which receipt and how I could use for this case?
I am sorry if this is trivial question but I am pretty new to prodigy .

Best regards,


hi @ula,

Thanks for your question and welcome to the Prodigy community :wave:

Try using --textcat-multilabel instead of --textcat. This thread explains why:

It explains why this is the case and also provides an alternative workaround near the end.

Hope this helps!