metadata

annotations_creators: []
language:
  - ro
language_creators:
  - machine-generated
license:
  - apache-2.0
multilinguality:
  - monolingual
pretty_name: BlackKakapo/t5-small-grammar-ro-root
size_categories:
  - 100K<n<1M
source_datasets:
  - original
tags: []
task_categories:
  - text2text-generation
task_ids: []

Romanian grammar

This model is trained on sentences with words in the root form. To bring the words to a syntactically correct form.

Prefix

grammar: text

How to use

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tokenizer = AutoTokenizer.from_pretrained("BlackKakapo/t5-small-grammar-ro-root")
model = AutoModelForSeq2SeqLM.from_pretrained("BlackKakapo/t5-small-grammar-ro-root")

Or

pip install happytransformer

from happytransformer import TTSettings, HappyTextToText

happy_tt_save = HappyTextToText('T5',r"BlackKakapo/t5-small-grammar-ro")
beam_settings =  TTSettings(num_beams=10, min_length=10, max_length=100)

Generate

sent = "As dori ca sa corecteze toate greșelile."
prefix = "grammar: "
example = prefix + sent

result = happy_tt_save.generate_text(example, args=beam_settings)

print(result.text)

Output

Aș dori ca să corecteze toate greșelile.