Create README.md

README.md

@@ -0,0 +1,39 @@
+# 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)