yaakov commited on
Commit
db53cfe
1 Parent(s): b031b03

Create Readme.md

Browse files
Files changed (1) hide show
  1. README.md +82 -0
README.md ADDED
@@ -0,0 +1,82 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - no-annotation
4
+ language_creators:
5
+ - crowdsourced
6
+ pretty_name: wikipedia-de-splits
7
+ paperswithcode_id: null
8
+ license:
9
+ - cc-by-sa-3.0
10
+ - gfdl
11
+ task_categories:
12
+ - text-generation
13
+ - fill-mask
14
+ task_ids:
15
+ - language-modeling
16
+ - masked-language-modeling
17
+ source_datasets:
18
+ - wikipedia
19
+ size_categories:
20
+ - n<1K
21
+ - 1K<n<10K
22
+ - 10K<n<100K
23
+ - 100K<n<1M
24
+ - 1M<n<10M
25
+ language:
26
+ - de
27
+ configs:
28
+ - "1"
29
+ - "2"
30
+ - "3"
31
+ - "4"
32
+ - "5"
33
+ - "6"
34
+ - "7"
35
+ - "8"
36
+ - "9"
37
+ - "10"
38
+ - "11"
39
+ - "12"
40
+ - "13"
41
+ - "14"
42
+ - "15"
43
+ - "16"
44
+ - "17"
45
+ - "18"
46
+ - "19"
47
+ - "20"
48
+ - "21"
49
+ - "all"
50
+ ---
51
+ # Dataset Card for yaakov/wikipedia-de-splits
52
+
53
+ ## Dataset Description
54
+ The only goal of this dataset is to have random German Wikipedia articles at
55
+ various dataset sizes: Small datasets for fast development and large datasets for statistically relevant measurements.
56
+
57
+ For this purpose, I loaded the 2665357 articles in the `test` set of the pre-processed German Wikipedia dump from 2022-03-01, randomly permuted the articles and created splits of sizes `2**n`: `1, 2, 4, 8, ...`. The split names are strings. The split `'all'` contains all 2665357 available articles.
58
+
59
+ ## Dataset creation
60
+ This dataset has been created with the following script:
61
+
62
+ !apt install git-lfs
63
+ !pip install -q transformers datasets
64
+
65
+ from huggingface_hub import notebook_login
66
+ notebook_login()
67
+
68
+ from datasets import load_dataset
69
+ wikipedia_de = load_dataset("wikipedia", "20220301.de")['train']
70
+
71
+ shuffled = wikipedia_de.shuffle(seed=42)
72
+
73
+ from datasets import DatasetDict
74
+ res = DatasetDict()
75
+
76
+ k, n = 0, 1
77
+ while n <= shuffled.num_rows:
78
+ res[str(k)] = shuffled.select(range(n))
79
+ k += 1; n *= 2
80
+ res['all'] = shuffled
81
+
82
+ res.push_to_hub('yaakov/wikipedia-de-splits')