vietgpt/

news_summarization_vi

Source: https://github.com/binhvq/news-corpus Language: Vietnamese Labeling: text-davinci-003 Num examples: 65,361 (train) 10,000 (test)

from datasets import load_dataset

load_dataset("tdtunlp/news_summarization_vi")

Format for Summarization task

def preprocess( sample, sep_key="<|endofprompt|>", end_key="<|endoftext|>", ): article = sample['content'] completion = sample['summary'] return {'text': """{article} TL;DR: {sep_key}… See the full description on the dataset page: https://huggingface.co/datasets/vietgpt/news_summarization_vi.