# Model description

- Morphosyntactic analyzer: Stanza
- Tagset: UD
- Embedding vectors: Fasttext (wiki)
- Dataset: PDB (http://git.nlp.ipipan.waw.pl/alina/PDBUD/tree/master/PDB-UD/PDB-UD)

# How to use

## Clone

```
git clone git@hf.co:ipipan/nlpre_stanza_ud_fasttext_pdb
```

## Load model

```
import stanza
lang = 'pl'
model_name = 'nlpre_stanza_ud_fasttext_pdb'
prefix = 'pdb1809'
config = \
  {
      # Comma-separated list of processors to use
      'processors': 'tokenize,mwt,pos,lemma',
      # Language code for the language to build the Pipeline in
      'lang': lang,
      # Processor-specific arguments are set with keys "{processor_name}_{argument_name}"
      # You only need model paths if you have a specific model outside of stanza_resources
      'tokenize_model_path': os.path.join(model_name, f'{lang}_{prefix}_tokenizer.pt'),
      'mwt_model_path': os.path.join(model_name, f'{lang}_{prefix}_mwt_expander.pt'),
      'pos_model_path': os.path.join(model_name, f'{lang}_{prefix}_tagger.pt'),
      'pos_pretrain_path': os.path.join(model_name, f'{lang}_{prefix}.pretrain.pt'),
      'lemma_model_path': os.path.join(model_name, f'{lang}_{prefix}_lemmatizer.pt'),
      # Use pretokenized text as input and disable tokenization
      'tokenize_pretokenized': True
  }
model = stanza.Pipeline(**config)