Coedit-large Open-source Text Editing Model - Free Deployment, Supports Various Text Revision Tasks

Coedit Large

Developed by grammarly

CoEdIT-Large is a text editing model fine-tuned from the google/flan-t5-large model, supporting various text revision tasks.

Large Language Model

Transformers

English#Instruction Fine-Tuned Text Editing #Multi-Task Text Revision #Grammar Correction and Optimization

Downloads 6,018

Release Time : 5/11/2023

Model Overview

This model is obtained by fine-tuning the flan-t5-large model on the CoEdIT dataset, primarily used for text editing and revision based on instructions.

Model Features

Multi-Task Text Editing

Supports various text editing tasks, including grammar correction, coherence improvement, simplification, rewriting, etc.

Instruction-Driven

Performs text revisions based on specific editing instructions provided by users.

High-Quality Revision

Demonstrates excellent performance in multiple text revision benchmarks.

Model Capabilities

Grammar correction

Text coherence improvement

Text simplification

Text rewriting

Formal expression

Neutral expression

Use Cases

Writing Assistance

Grammar Correction

Corrects grammatical errors in the text

Generates a grammatically correct version of the text

Text Coherence Improvement

Improves the coherence and fluency of the text

Generates more coherent text

Content Optimization

Text Simplification

Simplifies complex text into a more understandable version

Generates simpler and more comprehensible text

Style Conversion

Converts text to a more formal or neutral style

Generates text that matches the target style

🚀 Model Card for CoEdIT-Large

This model is fine - tuned from google/flan - t5 - large on the CoEdIT dataset, enabling various text editing tasks.

🚀 Quick Start

The models presented in our paper are available. You can use them for different text editing tasks as shown in the usage example below.

✨ Features

Multitask Text Editing: Can handle tasks such as grammar fixing, text coherence improvement, simplification, paraphrasing, formalization, and neutralization.
Fine - tuned from a Strong Base: Based on the google/flan - t5 - large model, fine - tuned on the CoEdIT dataset for better performance.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, T5ForConditionalGeneration

tokenizer = AutoTokenizer.from_pretrained("grammarly/coedit-large")
model = T5ForConditionalGeneration.from_pretrained("grammarly/coedit-large")
input_text = 'Fix grammatical errors in this sentence: When I grow up, I start to understand what he said is quite right.'
input_ids = tokenizer(input_text, return_tensors="pt").input_ids
outputs = model.generate(input_ids, max_length=256)
edited_text = tokenizer.decode(outputs[0], skip_special_tokens=True)

📚 Documentation

Model Details

Model Description

Property	Details
Language(s) (NLP)	English
Finetuned from model	google/flan - t5 - large

Model Sources

Property	Details
Repository	https://github.com/vipulraheja/coedit
Paper	https://arxiv.org/abs/2305.09857

Uses

Text Revision Task

Given an edit instruction and an original text, our model can generate the edited version of the text.

task_specs

Software

https://github.com/vipulraheja/coedit

Available Models

Model	Number of parameters
CoEdIT - large	770M
CoEdIT - xl	3B
CoEdIT - xxl	11B

Example Widgets

Task	Example Text
Fluency	Fix the grammar: When I grow up, I start to understand what he said is quite right.
Coherence	Make this text coherent: Their flight is weak. They run quickly through the tree canopy.
Simplification	Rewrite to make this easier to understand: A storm surge is what forecasters consider a hurricane's most treacherous aspect.
Paraphrase	Paraphrase this: Do you know where I was born?
Formalize	Write this more formally: omg i love that song im listening to it right now
Neutralize	Write in a more neutral way: The authors' exposé on nutrition studies.

Datasets and Metrics

Property	Details
Datasets	facebook/asset, wi_locness, GEM/wiki_auto_asset_turk, discofuse, zaemyung/IteraTeR_plus, jfleg, grammarly/coedit
Metrics	sari, bleu, accuracy

📄 License

This model is licensed under the CC - BY - NC - 4.0 license.

🔧 Technical Details

No technical details provided in the original document.

📄 Citation

BibTeX:

@article{raheja2023coedit,
      title={CoEdIT: Text Editing by Task - Specific Instruction Tuning}, 
      author={Vipul Raheja and Dhruv Kumar and Ryan Koo and Dongyeop Kang},
      year={2023},
      eprint={2305.09857},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

APA: Raheja, V., Kumar, D., Koo, R., & Kang, D. (2023). CoEdIT: Text Editing by Task - Specific Instruction Tuning. ArXiv. /abs/2305.09857

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