Datasets:
Download schema/record.json from PhysicalSI/ISB_GridNavLM_Dataset: direct link, hf CLI and curl.
- Browser
- Download file 1.73 kB
-
https://huggingface.co/datasets/PhysicalSI/ISB_GridNavLM_Dataset/resolve/main/schema/record.json
- Command line
-
hf download hf://datasets/PhysicalSI/ISB_GridNavLM_Dataset/schema/record.json
-
curl -L -o record.json https://huggingface.co/datasets/PhysicalSI/ISB_GridNavLM_Dataset/resolve/main/schema/record.json
1.73 kB
| { | |
| "schema_version": "navgridlm.hf-text-task/2", | |
| "encoding": "UTF-8 JSONL gzip", | |
| "input_columns": [ | |
| "text", | |
| "sections_json", | |
| "task_json" | |
| ], | |
| "answer_columns": [ | |
| "ground_truth_json", | |
| "shortest_length_path", | |
| "shortest_path_length_m" | |
| ], | |
| "group_splits_by": "source_scene_id", | |
| "split": "train | validation | test", | |
| "dynamic_json_columns": [ | |
| "sections_json", | |
| "task_json", | |
| "goals_json", | |
| "ground_truth_json", | |
| "metadata_json" | |
| ], | |
| "shared_assets": [ | |
| "map_ref", | |
| "robot_ref" | |
| ], | |
| "reference_schema": "navgridlm.text-task/2", | |
| "columns": { | |
| "schema_version": "string", | |
| "id": "string, unique task ID", | |
| "scenario_id": "string, shared by paired tasks", | |
| "source_scene_id": "string, group all variants/robots when splitting", | |
| "map_id": "string", | |
| "robot_id": "string", | |
| "task_type": "single_goal | ordered_goals | visit_all_goals", | |
| "goal_count": "integer: 1,2,3,5", | |
| "language": "string: vi", | |
| "split": "string: train | validation | test", | |
| "start_row": "integer", | |
| "start_column": "integer", | |
| "text": "string, full input context", | |
| "sections_json": "JSON string, selectable input text", | |
| "task_json": "JSON string, start/goals/order/budget/completion", | |
| "goals_json": "JSON string, goal IDs and cells", | |
| "ground_truth_json": "JSON string, exact route/actions/events/segments/metrics", | |
| "shortest_length_path": "integer, grid steps", | |
| "shortest_path_length_m": "number, meters", | |
| "metadata_json": "JSON string, generation provenance and answer-derived analysis", | |
| "map_ref": "relative map asset path", | |
| "robot_ref": "relative robot profile path" | |
| } | |
| } | |