A new version of Vietnamese Hate Speech Detection with target-oriented labels. Each comment can have multiple targets, each target has a hatred level indicating the hateful: CLEAN, OFFENSIVE, and HATE.
The targets are: Individuals, Groups, Religion/creed, Race/ethnicity, and Politics
Directly load and use the dataset from hugging face:
from datasets import… See the full description on the dataset page:
https://huggingface.co/datasets/sonlam1102/vithsd.
