In spaCy, there are indirect methods to retrieve frequency-related information for a word form or lemma based on the data the model was trained on, though spaCy doesn’t provide explicit frequency counts.
- SpaCy’s prob attribute in Lexeme gives a log probability based on the training corpus, which can be used as an indicator of general frequency.
- Vector norms can act as a rough frequency indicator but are less precise.
- For exact frequencies, consider combining spaCy with external frequency resources like wordfreq.