bkai-foundation-models/vi-alpaca

---
dataset_info:
features:
- name: instruction
dtype: string
- name: input
dtype: string
- name: output
dtype: string
splits:
- name: train
num_bytes: 51605940
num_examples: 50006
download_size: 26279260
dataset_size: 51605940
---
# 🇻🇳 Vietnamese Alpaca Dataset

This dataset is especially designed for Vietnamese based on the idea from [Stanford Alpaca](https://github.com/tatsu-lab/stanford_alpaca) and [Self-Instruct paper](https://arxiv.org/abs/2212.10560). The motivation behind the creation of this dataset stems from the hope to contribute high-quality dataset to Vietnamese commnunity to train language models.

To construct this dataset, we follow a two-step process:

- Step 1: Manually create Vietnamese seed tasks
We employ the methodology outlined in the [Self-Instruct paper](https://arxiv.org/abs/2212.10560) we meticulously curated a diverse set of seed tasks for the Vietnames with GPT-4 and hand-craft as well.

- Step 2: Instruction Generation
Building upon the manually created seed tasks, we employed the instruction generation process inspired by [Stanford Alpaca](https://github.com/tatsu-lab/stanford_alpaca). Using GPT-4, GPT-3.5 turbo, and GPT-3.5-instruct, we produced 50K instructions, employing diverse configurations to ensure a comprehensive and varied set of linguistic contexts.

### Please cite our manuscript if this dataset is used for your work
```
@article{duc2024towards,
title={Towards Comprehensive Vietnamese Retrieval-Augmented Generation and Large Language Models},
author={Nguyen Quang Duc, Le Hai Son, Nguyen Duc Nhan, Nguyen Dich Nhat Minh, Le Thanh Huong, Dinh Viet Sang},
journal={arXiv preprint arXiv:2403.01616},
year={2024}
}
```

原始信息汇总

🇻🇳 Vietnamese Alpaca Dataset

数据集信息

• 特征:
  • instruction: 类型为字符串
  • input: 类型为字符串
  • output: 类型为字符串
• 分割:
  • train: 字节数为51605940，样本数为50006
• 下载大小: 26279260字节
• 数据集大小: 51605940字节

数据集构建过程

1. 手动创建越南语种子任务:

   • 采用Self-Instruct paper中描述的方法，使用GPT-4和手工制作的方式精心筛选了一系列多样化的越南语种子任务。

2. 指令生成:

   • 基于手动创建的种子任务，采用Stanford Alpaca的指令生成过程。使用GPT-4、GPT-3.5 turbo和GPT-3.5-instruct生成了50K条指令，采用多样化的配置以确保语言环境的全面性和多样性。

引用

@article{duc2024towards, title={Towards Comprehensive Vietnamese Retrieval-Augmented Generation and Large Language Models}, author={Nguyen Quang Duc, Le Hai Son, Nguyen Duc Nhan, Nguyen Dich Nhat Minh, Le Thanh Huong, Dinh Viet Sang}, journal={arXiv preprint arXiv:2403.01616}, year={2024} }