| { | |
| "name": "syn4d", | |
| "manifests": [ | |
| "/workspace/dataset/Syn4D/v3/conditions/pairs.jsonl" | |
| ], | |
| "target_cache": "/workspace/dataset/Syn4D/v3/targets", | |
| "text_cache": "/workspace/dataset/Syn4D/v3/text", | |
| "path_map": {}, | |
| "conditions_code": "/workspace/code/Track4DGEN", | |
| "rgb_root": "/workspace/dataset/Syn4D/syn4d_tracks", | |
| "pairs": 12051, | |
| "clips": 5682, | |
| "chunk_frames": 33, | |
| "latent_frames": 5, | |
| "fps": 15, | |
| "target_minus_source_start": {}, | |
| "storage_format": "bf16", | |
| "metres": "Syn4D rendered depth + camera CSV (path B caches, scale 1.0)", | |
| "notes": "Syn4D (Syn4D/Syn4D @bb08d13f, CC BY 4.0) path B: tools/syn4d (ground-truth caches, selection dynamic_first_v3, boxes gravity_median_box_v1 on the ground-truth gravity, README \uc120\ubcc4)" | |
| } | |