Felix
commited on
Commit
•
2bcc84f
1
Parent(s):
e448142
Corrected the section on Data splits
Browse files
README.md
CHANGED
@@ -86,7 +86,7 @@ Swedish
|
|
86 |
|
87 |
### Data Splits
|
88 |
|
89 |
-
|
90 |
|
91 |
## Dataset Creation
|
92 |
|
|
|
86 |
|
87 |
### Data Splits
|
88 |
|
89 |
+
Most datasets have a train, dev and test split. However, there are a few (`supersim`, `sweanalogy` and `swesat-synonyms`) who only have a train and test split. The diagnostic tasks `swediagnostics` and `swewinogender` only have a test split, but they could be evaluated on models trained on `swenli` since they are also NLI-based.
|
90 |
|
91 |
## Dataset Creation
|
92 |
|