AptaBench Anonymous commited on
Update README.md
Browse files
README.md
CHANGED
|
@@ -44,10 +44,13 @@ Inactive labels are based on reported non-binding or low-affinity observations,
|
|
| 44 |
|
| 45 |
## Files
|
| 46 |
|
| 47 |
-
- `AptaBench_dataset.csv`: main aptamer–ligand interaction table.
|
| 48 |
-
- `
|
| 49 |
-
- `
|
| 50 |
-
- `
|
|
|
|
|
|
|
|
|
|
| 51 |
|
| 52 |
## Data fields
|
| 53 |
|
|
|
|
| 44 |
|
| 45 |
## Files
|
| 46 |
|
| 47 |
+
- `dataset/AptaBench_dataset.csv`: main aptamer–ligand interaction table in CSV format.
|
| 48 |
+
- `dataset/AptaBench_dataset.parquet`: parquet version of the dataset for efficient loading and processing.
|
| 49 |
+
- `dataset/splits/`: directory containing fixed leakage-aware evaluation splits.
|
| 50 |
+
- `stratified.json`: stratified in-distribution 5-fold split.
|
| 51 |
+
- `disjoint_molecule.json`: molecule-disjoint 5-fold split.
|
| 52 |
+
- `disjoint_aptamer.json`: aptamer-disjoint 5-fold split.
|
| 53 |
+
- `dataset/make_splits.py`: script used to generate evaluation splits.
|
| 54 |
|
| 55 |
## Data fields
|
| 56 |
|