Add Captioning task; reorder tasks REC, Captioning, TOG, TRG, OCR

Add raw JSON for Captioning task

Add nuscenes_captioning config

Sync README edits

Reorganize into nuscenes/<task>/ subfolders, clean up stale config metadata, use test split throughout

Move REC raw files into referring_expression/ subfolder

Remove stale pre-rename files (1600-1743)

Remove stale pre-rename files (1200-1600)

Remove stale pre-rename files (800-1200)

Remove stale pre-rename files (400-800)

Remove stale pre-rename files (0-400)

Move Text-Region Grounding config to test split

Move Text-Object Grounding config to test split

Move OCR config to test split

Rename REC config, move to test split

Add raw JSON + bbox images for OCR, Text-Object Grounding, Text-Region Grounding

Add nuscenes_text_region_grounding config

Add nuscenes_text_object_grounding config

Add nuscenes_ocr config

Add nuscenes config as Parquet with embedded images for the Dataset Viewer

Per-sample bbox images: only the referenced object + text box(es) for each sample

Remove old per-image bbox files (replaced by per-sample bbox images)

Add nuscenes config as Parquet with embedded images for the Dataset Viewer

Sync trimmed README

Update README to align with the Read the Object paper

Add nuScenes source: images, bbox images, and REC annotations

initial commit