How can I list all installed models? I want to know which models I can use for text classification.
I hope I undderstand your question correctly! If you’re using spaCy, you can run
python -m spacy validate to see all available models in the environment (and whether they’re up to date).
For a list of all available pre-trained statistical models, see this page:
Even if spaCy doesn’t ship with a pre-trained statistical model, you can still export a blank model for one of the available languages and add a text classification component to it. To export a blank model, you can run the following:
import spacy nlp = spacy.blank('ar') # for example, a blank Arabic model nlp.to_disk('/path/to/model')
Or, as a handy one-liner:
python -c "import spacy; spacy.blank('ar').to_disk('/path/to/model')"
In Prodigy, you can then use
/path/to/model as the model you load into the recipe and collect annotations.
To use a different library and/or model implementation, you can also write a custom recipe so you can annotate with a model in the loop. Here’s a code example that shows how this could look for text classification.
Perfect. That worked! thanks!