Are you a vendor experienced in data collection? We have exciting opportunities for dataset generation in the following categories:

1. AI Agent Non-Open Source COT/TOT/GOT Dataset

  • Description: Decompose tasks into multiple steps with corresponding thoughts and behaviors for each step to obtain results. Convert PDF to text, ensuring tables, codes, and formulas are in standard format.
  • Volume: 1 Million pieces
  • Language: English
  • Format: JSON
  • Delivery: FTP, hard disk, network; batch delivery.

2. High-Quality Video Description Dataset

  • Description: Collect genuine video works published globally, with author descriptions and tags.
  • Volume: 10 Million sets (7 million in English, 3 million in Chinese)
  • Video Resolution: 1080p and above
  • Languages: English, Chinese
  • Format: .mp4, .mov, .avi; XLSX for annotations.

3. High School Level Examination Questions

  • Description: Off-the-shelf or custom examination questions.
  • Languages: English, Chinese
  • Volume: To be discussed based on availability.

Interested vendors can propose costs and other details apply via link