ando55 commited on
Commit
d71aa73
·
verified ·
1 Parent(s): 4067dc6

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -0
README.md CHANGED
@@ -249,6 +249,8 @@ While the parent dataset (`ando55/WikiSQE`) contains **every** sentence flagged
249
  * **27 dataset *groups*** (20 frequent quality labels + 5 Quality type categories + 2 Coarse groups)
250
  * **3 standard splits per group** (`train`, `val`, `test`) – for example `citation/train`, `citation/val`, …
251
 
 
 
252
  > **Need the full dump?** Head to [https://huggingface.co/datasets/ando55/WikiSQE](https://huggingface.co/datasets/ando55/WikiSQE).
253
 
254
  ---
 
249
  * **27 dataset *groups*** (20 frequent quality labels + 5 Quality type categories + 2 Coarse groups)
250
  * **3 standard splits per group** (`train`, `val`, `test`) – for example `citation/train`, `citation/val`, …
251
 
252
+ Each split blends **labeled** and **unlabeled** sentences at a **1 : 1 ratio** to support semi-supervised and positive/negative training paradigms.
253
+
254
  > **Need the full dump?** Head to [https://huggingface.co/datasets/ando55/WikiSQE](https://huggingface.co/datasets/ando55/WikiSQE).
255
 
256
  ---