metadata: migrate score_type -> score_pipeline (atomic stage contract; see mm-eval scorer docs/en/SCORING.md)

format v3: flat sample layout (top-level answer/question_type/pass-throughs), task_type v3 vocabulary, judge info consolidated in metadata score block. See dataset-upload skill references/mmeval-format.md and field-census-2026-07-08.md.

metadata: 2026-07-07 revised taxonomy (task_type format vocabulary multiple_choice/yes_no/short_answer/free_form; score_type incl. none; re-verified protocol evidence; repaired prompt provenance)

metadata: add score_type/score_params/score_protocol (official grading protocol taxonomy); task_type is now format-only (judge_scored migrated to vqa + llm_judge). See dataset-upload skill references/metadata-json.md.

Tighten prompt_template_source.notes: explicit 5-tier classification per 2026-06-01 priority rules

Refine prompt_template_source: cite verifiable official sources per 2026-05-30 deep audit

Populate prompt_template_source for all subsets per 2026-05-30 audit

2026-05-28 prompt-template audit: align to authoritative source

fix metadata.json source.url key test->train to match actual split

Upload folder using huggingface_hub

Upload captioning/data.json with huggingface_hub

Upload caption_matching/data.json with huggingface_hub

Upload yes_no/data.json with huggingface_hub

Upload multi_choice/data.json with huggingface_hub

Upload metadata.json with huggingface_hub

initial commit