Labels

• Agentic

  For Agentic AI, e.g., a tool-based agent.
• Bench

  Benchmark
• BLIP2

  Generated by BLIP2
• Cap.

  Image Captioning
• Claude-3.5
• Cls.

  Scene Classification
• DCap.

  Detailed Captioning
• DeepSeek-V3

  Generated by DeepSeek-V3
• Doubao-1.5

  Generated by Doubao-1.5
• Doubao-1.6
• GCap.

  Grounded Captioning
• Gemini-1.5-flash
• Gemini-2.5
• Gemini-2.5-pro
• Gemini-Vision

  Generated by Gemini-Vision
• Geo-Loc.

  Geo-Localization
• good first dataset

  Good for newcomers
• GPT-3.5

  Generated by GPT-3.5
• GPT-4

  Generated by GPT-4
• GPT-4.1

  Generated by GPT-4.1
• GPT-4o

  Generated by GPT-4o
• GPT-4o-mini
• GPT-4V

  Generated by GPT-4V
• GPT-5-mini
• GREC

  Generalized Referring Expression Comprehension
• GRES

  Generalized Referring Expression Segmentation
• GVQA

  Grounding Visual Question Answering
• HBB

  Horizontal Bounding Box (Annotation of VG)
• Human

  Human Annotation
• InternLM-XComposer

  Generated by InternLM-XComposer2