Commit History

  1. License metadata: cc-by-nc-sa-4.0 (matches source datasets and card text)

    Verified signature
    790f9d4

  1. README: restore original voice, tasks overview covers both tasks, clarify countries and citations

    Verified signature
    23e699c
  2. Restructure README: Task 1 and Task 2 sections, GitHub starter kits; citations already complete

    Verified signature
    9ab8353
  3. Task 2: single config with one merged jsonl per split

    Verified signature
    0def513
  4. Task 2: one merged jsonl per split (captions + gold + image paths); drop baselines and per-file tsvs

    Verified signature
    e6b133b
  5. Consolidated repo: rename, add Task 2 section and configs, mark all gold labels released

    Verified signature
    f3224fc
  6. Add Task 2 (CRAI-Bench): train/dev/test with captions, human gold labels, reference and generated images, organizer baselines

    Verified signature
    2ebec70
  7. Release test gold labels; add country/category/subcategory metadata to devtest and test

    Verified signature
    0909d4a

  1. Add gold labels to devtest split

    Verified signature
    f25d266
  2. Declare test split in dataset configs

    Verified signature
    5001fb3
  3. Release blind test split (task1a, task1b)

    Verified signature
    6362d41
  4. Add media (part 8)

    Verified signature
    4fc81f7
  5. Add media (part 7)

    Verified signature
    0ffa110
  6. Add media (part 6)

    Verified signature
    ed8b71e
  7. Add media (part 5)

    Verified signature
    4a55572
  8. Add media (part 4)

    Verified signature
    9d8e9ab
  9. Add media (part 3)

    Verified signature
    5f18b94
  10. Add media (part 2)

    Verified signature
    8903f38
  11. Add media

    Verified signature
    fdd65ec

  1. updated bib

    Verified signature
    4e0ecbd

  1. Update Slack invite link

    Verified signature
    757bf95

  1. Fix dataset card: restore task1b configs (was broken by stale task1c paths); keep Slack link

    Verified signature
    b7fab10

  1. Remove em-dashes from dataset card

    Verified signature
    2fe0260
  2. Dataset card: replace Google registration form with Slack channel

    Verified signature
    8e3806a

  1. Codabench links -> re-uploaded comps (1a 17048/17049, 1b 17050/17051)

    Verified signature
    cf5764f
  2. Task 1b: ranking metric -> Contrastive Instability; CFHR/Q+/Q- diagnostics; Codabench 17045/17046; greedy-decoding note

    Verified signature
    fdaa4de

  1. Update README.md

    Verified signature
    0c4cd2f
  2. Update README.md

    Verified signature
    8f8c74e
  3. Update README.md

    Verified signature
    de5e45c

  1. Update Task 1b Codabench links to new competitions (17022 en / 17021 msa)

    Verified signature
    b051d5d
  2. Remove old task1c/ folder (renamed to task1b)

    Verified signature
    cb73862
  3. Add task1b/ data (copy of task1c) to match card configs

    Verified signature
    752ead8
  4. Update README.md

    Verified signature
    604e8f5
  5. updated subtask id

    Verified signature
    20c41b3
  6. Card: call the Fanar cascade an example, not a baseline (consistency with task pages)

    Verified signature
    75abca9
  7. Clean dataset card: remove em-dashes (no content or number changes)

    Verified signature
    cf66ad1

  1. ImageEval 2026 Task 1 (Ayn-VQA) dataset

    ec05656