Coedit Xxl

🚀 Model Card for CoEdIT-xxl

This model card provides detailed information about the CoEdIT-xxl model, which is fine - tuned from the google/flan - t5 - xxl model on the CoEdIT dataset. It offers insights into the model's details, usage, and citation information.

🚀 Quick Start

The CoEdIT-xxl model is readily available for use. You can start using it by following the code example below:

from transformers import AutoTokenizer, T5ForConditionalGeneration

tokenizer = AutoTokenizer.from_pretrained("grammarly/coedit-xxl")
model = T5ForConditionalGeneration.from_pretrained("grammarly/coedit-xxl")
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)

✨ Features

• Text Revision: Given an edit instruction and an original text, the model can generate the edited version of the text.
• Multiple Model Sizes: Available in different sizes (CoEdIT-large, CoEdIT-xl, CoEdIT-xxl) with varying numbers of parameters.

📚 Documentation

📦 Installation

The model can be installed using the transformers library. Ensure you have it installed in your Python environment.

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, T5ForConditionalGeneration

tokenizer = AutoTokenizer.from_pretrained("grammarly/coedit-xxl")
model = T5ForConditionalGeneration.from_pretrained("grammarly/coedit-xxl")
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)

🔧 Technical Details

Model Details

Model Sizes

📄 License

This model is licensed under the cc - by - nc - 4.0 license.

📦 Datasets

The model is trained on the following datasets:

• facebook/asset
• wi_locness
• GEM/wiki_auto_asset_turk
• discofuse
• zaemyung/IteraTeR_plus
• jfleg
• grammarly/coedit

📊 Metrics

The performance of the model is evaluated using the following metrics:

• sari
• bleu
• accuracy

📖 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