Skip to content

Latest commit

 

History

History
31 lines (28 loc) · 795 Bytes

File metadata and controls

31 lines (28 loc) · 795 Bytes

Web-generate-code-and-Web-judge-code

Description: Repository for generating and evaluating code for various types of tasks

Repository Structure

Generate_code

  • text-input
    • generate(single-round/multi-round)
    • edit
    • repair
  • image-input
    • generate(single image/multiple images)
    • edit(single image/multiple images)
    • repair(single image/multiple images)
  • video-input
    • generate

Data Example(题目以及checklist)

  • text-input
    • generate(single-round/multi-round)
    • edit
    • repair
  • image-input
    • generate(single image/multiple images)
    • edit(single image/multiple images)
    • repair(single image/multiple images)
  • video-input
    • generate

Judge_code

  • LLM-as-a-judge(visual: 美观/instruction following)
  • agent-as-a-judge(web-agent/llm看视频)