prediction.md

Prediction

COMBO as a Python library

The pre-trained models can be automatically downloaded with the from_pretrained method. Select a model name from the lists: UD-trained COMBO models and enhanced COMBO models, and pass it as an argument of from_pretrained.

from combo.predict import COMBO

nlp = COMBO.from_pretrained("polish-herbert-base-ud29")
sentence = nlp("Sentence to parse.")

You can also load your own COMBO model:

from combo.predict import COMBO

model_path = "your_model.tar.gz"
nlp = COMBO.from_pretrained(model_path)
sentence = nlp("Sentence to parse.")

COMBO allows to enter presegmented sentences (or texts:

from combo.predict import COMBO

model_path = "your_model.tar.gz"
nlp = COMBO.from_pretrained(model_path)
tokenized_sentence = ["Sentence", "to", "parse", "."]
sentence = nlp([tokenized_sentence])

You can use COMBO with the LAMBO tokeniser (Note: installing LAMBO is necessary, see LAMBO installation ).

# Import COMBO and lambo
from combo.predict import COMBO
from combo.utils import lambo_tokenizer

# Download models
nlp = COMBO.from_pretrained("english-bert-base-ud29",tokenizer=lambo_tokenizer.LamboTokenizer("en"))
sentences = nlp("This is the first sentence. This is the second sentence to parse.")

COMBO as a command-line interface

CoNLL-U file prediction:

Input and output are both in *.conllu format.

combo --mode predict --model_path your_model_tar_gz --input_file your_conllu_file --output_file your_output_file --silent

Raw text prediction:

Works for models where input was text-based only.

Input: one sentence per line.

Output: List of token jsons.

combo --mode predict --model_path your_model_tar_gz --input_file your_text_file --output_file your_output_file --silent --noconllu_format

Console prediction:

Works for models where input was text-based only.

Interactive testing in console (load model and just type sentence in console).

combo --mode predict --model_path your_model_tar_gz --input_file "-" --nosilent

Advanced

There are 2 tokenizers: whitespace and spacy-based (en_core_web_sm model).

Use either --predictor_name combo or --predictor_name combo-spacy (default tokenizer).