AMRBART Large V2

🚀 AMRBART

AMRBART is a pre - trained semantic parser that can convert a sentence into an abstract meaning graph. It offers a powerful solution for semantic analysis, enabling more accurate understanding and processing of natural language.

🚀 Quick Start

The model is available at huggingface. You can initialize an AMR parsing model in PyTorch as follows:

from transformers import BartForConditionalGeneration
from model_interface.tokenization_bart import AMRBartTokenizer      # We use our own tokenizer to process AMRs

model = BartForConditionalGeneration.from_pretrained("xfbai/AMRBART-large-finetuned-AMR3.0-AMRParsing-v2")
tokenizer = AMRBartTokenizer.from_pretrained("xfbai/AMRBART-large-finetuned-AMR3.0-AMRParsing-v2")

✨ Features

• State - of - the - art Performance: Achieves top results in AMR - to - text generation and AMR parsing on multiple datasets, as shown by the badges below:
  • 
  • 
  • 
  • 
• Model Updates:
  • (2022/12/10) Fixed max_length bugs in AMR parsing and updated results.
  • (2022/10/16) Released the AMRBART - v2 model which is simpler, faster, and stronger.

📦 Installation

Requirements

• python 3.8
• pytorch 1.8
• transformers 4.21.3
• datasets 2.4.0
• Tesla V100 or A100

We recommend using conda to manage virtual environments:

conda env update --name <env> --file requirements.yml

📚 Documentation

Data Processing

You can download the AMR corpora at LDC. Please follow this repository to preprocess AMR graphs:

bash run-process-acl2022.sh

📄 License

This project is licensed under the MIT license.

📖 Related Paper

You may find our paper here (Arxiv). The original implementation is available here