Problems with multi-label

[p0wned17]

Traceback (most recent call last):
  File "/home/powned/anaconda3/envs/venv/lib/python3.9/site-packages/spacy/training/loop.py", line 281, in evaluate
    scores = nlp.evaluate(dev_corpus(nlp))
  File "/home/powned/anaconda3/envs/venv/lib/python3.9/site-packages/spacy/language.py", line 1427, in evaluate
    results = scorer.score(examples)
  File "/home/powned/anaconda3/envs/venv/lib/python3.9/site-packages/spacy/scorer.py", line 137, in score
    scores.update(component.score(examples, **self.cfg))
  File "spacy/pipeline/pipe.pyx", line 94, in spacy.pipeline.pipe.Pipe.score
  File "/home/powned/anaconda3/envs/venv/lib/python3.9/site-packages/spacy/pipeline/textcat_multilabel.py", line 115, in textcat_multilabel_score
    return Scorer.score_cats(
  File "/home/powned/anaconda3/envs/venv/lib/python3.9/site-packages/spacy/scorer.py", line 492, in score_cats
    auc_per_type[label].score_set(pred_score, gold_score)
KeyError: 'balance, resource'

But in config exclusive_classes = False

[polm]

When reporting problems, it's hard to help if all we get is an error without some background about what you're trying to do or what led up to the issue.

Based on the information provided here, it looks like the issue is that multiple categories were set incorrectly, resulting in a label in the gold data that isn't in training. I suspect something like the "incorrect" example below happened:

# correct
doc.cats["balance"] = 1.0
doc.cats["resource"] = 1.0
# incorrect
doc.cats["balance, resource"] = 1.0

Can you check if that's the case? If that's not the case, can you share some more details of what you're trying to do, along with the output of spacy debug data?