mathiascreutz
commited on
Commit
•
4228fac
1
Parent(s):
bc72843
Minor modifications
Browse files
README.md
CHANGED
@@ -129,7 +129,12 @@ The entries in the training sets have been ranked automatically by how
|
|
129 |
likely they are paraphrases, best first, worst last. The quality
|
130 |
parameter indicates the estimated proportion (in percent) of true
|
131 |
paraphrases in the set. Allowed quality values range between 60 and
|
132 |
-
100, in increments of 5 (60, 65, 70,
|
|
|
|
|
|
|
|
|
|
|
133 |
|
134 |
|
135 |
### Data Instances
|
|
|
129 |
likely they are paraphrases, best first, worst last. The quality
|
130 |
parameter indicates the estimated proportion (in percent) of true
|
131 |
paraphrases in the set. Allowed quality values range between 60 and
|
132 |
+
100, in increments of 5 (60, 65, 70, ..., 100). A value of 60 means
|
133 |
+
that 60% of the sentence pairs in the downloaded set are estimated to
|
134 |
+
be true paraphrases (and the remaining 40% are not). A higher value
|
135 |
+
produces a smaller but cleaner set. The smaller sets are subsets of
|
136 |
+
the larger sets, such that the `quality=95`set is a subset of
|
137 |
+
`quality=90`, which is a subset of `quality=85`, and so on.
|
138 |
|
139 |
|
140 |
### Data Instances
|