{
  "workspace": "/home/hm/Projects/Tests",
  "downloaded_at": "2026-09-02",
  "datasets": [
    {
      "name": "HumanEval",
      "path": "datasets/humaneval",
      "purpose": "code generation",
      "split": "test",
      "items": 164,
      "license": "MIT",
      "source": "https://huggingface.co/datasets/openai/openai_humaneval"
    },
    {
      "name": "MBPP",
      "path": "datasets/mbpp",
      "purpose": "code generation",
      "split": "test + validation",
      "items": 590,
      "license": "CC BY 4.0",
      "source": "https://huggingface.co/datasets/google-research-datasets/mbpp"
    },
    {
      "name": "MMLU-Pro",
      "path": "datasets/mmlu_pro",
      "purpose": "general reasoning",
      "split": "test + validation",
      "items": 12102,
      "license": "MIT",
      "source": "https://huggingface.co/datasets/TIGER-Lab/MMLU-Pro"
    },
    {
      "name": "GenEval",
      "path": "tools/geneval",
      "purpose": "text-to-image compositional generation",
      "split": "official prompts",
      "items": 553,
      "license": "MIT repository release",
      "source": "https://github.com/djghosh13/geneval"
    },
    {
      "name": "T2I-CompBench",
      "path": "tools/t2i_compbench",
      "purpose": "text-to-image compositional generation",
      "split": "validation",
      "items": 2400,
      "license": "MIT repository/data mirror",
      "source": "https://github.com/Karine-Huang/T2I-CompBench"
    },
    {
      "name": "AudioCaps",
      "path": "datasets/audiocaps",
      "purpose": "text-to-audio prompt/reference metadata",
      "split": "test metadata",
      "items": 4875,
      "license": "MIT-tagged metadata; source media terms apply",
      "source": "https://huggingface.co/datasets/d0rj/audiocaps"
    },
    {
      "name": "Clotho",
      "path": "datasets/clotho",
      "purpose": "text-to-audio prompt/reference evaluation",
      "split": "test audio",
      "items": 1045,
      "license": "See included upstream README and source terms",
      "source": "https://huggingface.co/datasets/gijs/clotho"
    }
  ],
  "notes": [
    "Only test/validation material was downloaded; training splits were omitted.",
    "Image/audio generation outputs and evaluator model weights are not part of these dataset sizes.",
    "AudioCaps metadata does not include redistributable copies of the original YouTube clips."
  ]
}