Spivavtor-Large: An Open-Source Ukrainian Text Editing Model - Free Text Rewriting and Grammar Correction

Spivavtor Large

Developed by grammarly

Spivavtor-Large is an instruction fine-tuned Ukrainian text editing model focused on tasks such as text rewriting, simplification, grammar correction, and coherence optimization.

Large Language Model

Transformers

Other#Ukrainian text editing #Instruction fine-tuning #Grammar correction

Downloads 169

Release Time : 3/29/2024

Model Overview

This model is obtained by instruction fine-tuning the bigscience/mt0-large model with the Spivavtor dataset, primarily used for Ukrainian text editing tasks.

Model Features

Ukrainian text editing

Specially optimized for Ukrainian, supporting various text editing tasks.

Instruction fine-tuning

Achieves multiple text editing functions through instruction fine-tuning, such as rewriting, simplification, and grammar correction.

Multitasking support

Supports various text editing tasks, including rewriting, simplification, grammar correction, and coherence optimization.

Model Capabilities

Text rewriting

Sentence simplification

Grammar correction

Coherence optimization

Use Cases

Text editing

Sentence rewriting

Rewrite sentences to improve expression or adapt to different contexts.

Generates sentences with the same meaning but different expressions.

Sentence simplification

Simplify complex sentences to make them easier to understand.

Generates more concise and understandable sentences.

Grammar correction

Correct grammatical errors in sentences.

Generates grammatically correct sentences.

Coherence optimization

Optimize the coherence and fluency of sentences.

Generates more coherent and fluent sentences.

🚀 Model Card for Spivavtor-Large

This model is an instruction-tuned version of the bigscience/mt0-large model on the Spivavtor dataset. All details regarding the dataset and the fine-tuning process can be found in our paper.

Paper: Spivavtor: An Instruction Tuned Ukrainian Text Editing Model Authors: Aman Saini, Artem Chernodub, Vipul Raheja, Vivek Kulkarni

✨ Features

License: cc-by-4.0
Datasets: grammarly/spivavtor
Inference Parameters:
- max_new_tokens: 250

Usage Examples

Paraphrasing example: "Перефразуйте речення: Який найкращий комплiмент, який ти отримував вiд будь-кого?"
Simplification example: "Спростiть речення: Там він помер через шість тижнів, 13 січня 888 року."
GEC example: "Виправте граматику в цьому реченнi: Дякую за інформацію! ми з Надією саме вийшли з дому"
Coherence example: "Виправте зв’язнiсть в реченнi: Лінч досі відмовляється розповідати про сумнозвісну травневу дорожньо-транспортну пригоду, коли він збив жінку-пішохода в районі нічного клубу Баффало та відлетів. Той факт, що Лінч взагалі говорив, заслуговує на увагу в цих краях."

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("grammarly/spivavtor-large")
model = AutoModelForSeq2SeqLM.from_pretrained("grammarly/spivavtor-large")

# Paraphrase the sentence: What is the greatest compliment that you ever received from anyone?
input_text = 'Перефразуйте речення: Який найкращий комплімент, який ти отримував від будь-кого?'

inputs = tokenizer.encode(input_text, return_tensors="pt")
output = model.generate(inputs, max_length=256)
output_text = tokenizer.decode(outputs[0], skip_special_tokens=True)

📚 Documentation

Model Details

Property	Details
Language	Ukrainian
Finetuned from model	bigscience/mt0-large

Available Models

Model	Number of parameters	Reference name in Paper
Spivavtor-large	1.2B	SPIVAVTOR-MT0-LARGE
Spivavtor-xxl	13B	SPIVAVTOR-AYA-101

📄 License

The model is released under the cc-by-4.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご