Datasets:
ArXiv:
License:
Update README.md
Browse files
README.md
CHANGED
@@ -8,7 +8,9 @@ This dataset includes two human-annotated benchmarks: Objaverse-General (of 100
|
|
8 |
Inside both ```objaverse-general``` and ```objaverse-shapanetepart``` directories, the benchmark has the following directory structure:
|
9 |
Each object is stored in a folder with the structure below:
|
10 |
|
|
|
11 |
{object_name}_{objaverse_uid}/
|
12 |
βββ label_map.json # Maps part IDs to natural language part names
|
13 |
βββ labels.npy # Per-point label assignment (0 means no label)
|
14 |
βββ points5000.pcd # Object point cloud with 5000 points
|
|
|
|
8 |
Inside both ```objaverse-general``` and ```objaverse-shapanetepart``` directories, the benchmark has the following directory structure:
|
9 |
Each object is stored in a folder with the structure below:
|
10 |
|
11 |
+
```
|
12 |
{object_name}_{objaverse_uid}/
|
13 |
βββ label_map.json # Maps part IDs to natural language part names
|
14 |
βββ labels.npy # Per-point label assignment (0 means no label)
|
15 |
βββ points5000.pcd # Object point cloud with 5000 points
|
16 |
+
```
|