Commit ·
3a17f66
1
Parent(s): cf2bad6
Update README.md
Browse files
README.md
CHANGED
|
@@ -50,8 +50,19 @@ An example of A-FB15k237:
|
|
| 50 |
|
| 51 |
### Data Fields
|
| 52 |
|
| 53 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
|
|
|
|
|
|
|
|
|
|
| 55 |
### Data Splits
|
| 56 |
|
| 57 |
<table>
|
|
|
|
| 50 |
|
| 51 |
### Data Fields
|
| 52 |
|
| 53 |
+
The data fields are the same among all splits.
|
| 54 |
+
For EDIT sub-task:
|
| 55 |
+
|
| 56 |
+
- ori: the fact in the pre-train dataset.
|
| 57 |
+
- cor: corrupted triple.
|
| 58 |
+
- process: the triple after replacing the wrong entity with the [MASK] token.
|
| 59 |
+
- label: a classification label, the scope is the entire set of entities.
|
| 60 |
+
|
| 61 |
+
For ADD sub-task:
|
| 62 |
|
| 63 |
+
- triples: the knowledge that needs to be injected into the model.
|
| 64 |
+
- label: a classification label, the scope is the entire set of entities.
|
| 65 |
+
- head: the head or tail entity that does not appear in pre-train.
|
| 66 |
### Data Splits
|
| 67 |
|
| 68 |
<table>
|