Finetune model for Summarization

This project is a system for summarizing the context into a short sentence. In this project, i use model BARTpho and ViT5 for summarizing.

Source code

git clone https://github.com/TrongNV2003/Summarization-Newspaper.git

Usage

First, you need to install requirement pakage:

pip install -r requirements.txt

Newspaper dataset

You need to get summarization newspaper dataset. Get from HuggingFace.

Or clone from huggingface:

git clone https://huggingface.co/datasets/Trongdz/summarization-dataset

(Dataset should put inside source code folder path)

Fine-tune model

For fine-tune model, run:

python training/train_qa.py

Or can get pre-trained model: -BARTpho: HuggingFace -ViT5: HuggingFace

Test model

python training/test_qa.py

Evaluate model

python rouge_score.py

Run model

You can change any context in "data/context.txt" for summarizing, then run:

python run_summa.py