rafmacalaba commited on
Commit
f52d9db
·
verified ·
1 Parent(s): f89907e

re-chunk to <=384-token windows

Browse files
Files changed (4) hide show
  1. README.md +24 -0
  2. holdout.jsonl +0 -0
  3. train.jsonl +0 -0
  4. validation.jsonl +0 -0
README.md ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-4.0
3
+ task_categories:
4
+ - token-classification
5
+ tags:
6
+ - ner
7
+ - gliner2
8
+ - data-use
9
+ configs:
10
+ - config_name: default
11
+ data_files:
12
+ - split: train
13
+ path: train.jsonl
14
+ - split: validation
15
+ path: validation.jsonl
16
+ - split: holdout
17
+ path: holdout.jsonl
18
+ ---
19
+
20
+ # datause-dataset (re-chunked)
21
+
22
+ Re-chunk of `ai4data/datause-dataset` into <=384-token windows (max_tokens=384, overlap=50) so GLiNER's window is never truncated during evaluation.
23
+
24
+ `{"train": {"orig": 1779, "rechunked": 1822}, "validation": {"orig": 415, "rechunked": 458}, "holdout": {"orig": 1149, "rechunked": 1166}}`
holdout.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
validation.jsonl ADDED
The diff for this file is too large to render. See raw diff