readme: minor fixes
Browse files
README.md
CHANGED
@@ -24,7 +24,7 @@ The original training dataset contains 31,756 annotated tweets. Our new training
|
|
24 |
|
25 |
The test dataset remains unmodified and contains 3,528 tweets.
|
26 |
|
27 |
-
Twitter user mentions were substituted by
|
28 |
|
29 |
Each instance contains up to 1 labels corresponding to one of the following sub-task:
|
30 |
|
@@ -42,7 +42,7 @@ The dataset is based on Turkish.
|
|
42 |
|
43 |
### Data Instances
|
44 |
|
45 |
-
A binary dataset with with (NOT) Not Offensive and (OFF) Offensive tweets.
|
46 |
|
47 |
# Acknowledgements
|
48 |
|
|
|
24 |
|
25 |
The test dataset remains unmodified and contains 3,528 tweets.
|
26 |
|
27 |
+
Twitter user mentions were substituted by `@USER` and URLs have been substitute by `URL`.
|
28 |
|
29 |
Each instance contains up to 1 labels corresponding to one of the following sub-task:
|
30 |
|
|
|
42 |
|
43 |
### Data Instances
|
44 |
|
45 |
+
A binary dataset with with (`NOT`) Not Offensive and (`OFF`) Offensive tweets.
|
46 |
|
47 |
# Acknowledgements
|
48 |
|