Hindi-Indonesian Safety Data Annotation

Hindi and Indonesian Safety Hardening and Safety Dataset - Annotation
1. Annotation Requirement Description
This annotation task aims to construct safety datasets for Hindi and Indonesian through manual annotation.
1.1 Basic Task Information
Task Summary: Annotate five types of raw data (sensitive words, text samples, image samples, "image-text" pairs, "video-text" pairs) in Hindi and Indonesian according to requirements.
Deliverable Types and Formats:
a. Sensitive Words: Words, phrases. Delivered in Excel and JSONL formats only.
b. Text Samples: Sentences, paragraphs. Delivered in Excel and JSONL formats only.
c. Image Samples: Images in JPG or PNG format, stored in folders. Deliver Excel, JSONL, and corresponding attachment folders.
d. "Image-Text" Pairs: Consist of an image and matching text, stored in folders. Deliver Excel, JSONL, and corresponding attachment folders.
e. "Video-Text" Pairs: Consist of a video and matching text, no specific video format requirement, stored in folders. Deliver Excel, JSONL, and corresponding attachment folders.
Languages: 2 languages: Hindi and Indonesian.