humanizer-data-vi / README.md
toandev's picture
Add Vietnamese Humanizer Dataset
1bc9f0d
|
Raw History Blame Contribute Delete
1.33 kB
---
language:
- vi
license: cc-by-nc-4.0
configs:
- config_name: default
data_files:
- split: train
path: data/train-00000-of-00001.parquet
- split: validation
path: data/validation-00000-of-00001.parquet
- split: test
path: data/test-00000-of-00001.parquet
---
# Vietnamese Humanizer Dataset
SFT data for natural Vietnamese rewriting: an AI rewrite is the input and the original news passage is the target (`messages`).
- **2,677 pairs:** 2,151 train · 267 validation · 259 test, in Parquet format.
- **Generation and evaluation model:** `gpt-6.1-sol`.
- **Recorded API usage:** 8,916,167 tokens — 6,578,507 input and 2,337,660 output, including recorded trial calls.
- **Estimated cost:** approximately **USD 36.53** at [standard token prices](https://developers.openai.com/api/docs/models/gpt-6.1-sol).
- **Created by:** [toandev](https://huggingface.co/toandev).
Thanks to **Underthesea** for [UVN-1](https://huggingface.co/datasets/undertheseanlp/UVN-1), the source of news articles published in 2025–2026. Dataset construction was inspired by [jialinyyzz/humanizer-data](https://huggingface.co/datasets/jialinyyzz/humanizer-data); no samples from that dataset were used.
**CC BY-NC 4.0 — non-commercial use.** Automatically filtered; not fully human-reviewed.