Datasets:

Modalities:
Text
Formats:
csv
Languages:
English
ArXiv:
Libraries:
Datasets
pandas
License:
adamnarozniak commited on
Commit
abf98c4
·
verified ·
1 Parent(s): 4e428b1

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +129 -0
README.md CHANGED
@@ -1,3 +1,132 @@
1
  ---
2
  license: bsd-2-clause
 
 
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: bsd-2-clause
3
+ task_categories:
4
+ - text-generation
5
+ language:
6
+ - en
7
+ size_categories:
8
+ - 1M<n<10M
9
+ configs:
10
+ - config_name: default
11
+ data_files:
12
+ - split: train
13
+ path: "shakespeare.csv"
14
  ---
15
+ # Dataset Card for Dataset Name
16
+
17
+ This dataset is a part of the [LEAF](https://leaf.cmu.edu/) benchmark.
18
+ The Shakespeare dataset is built from [The Complete Works of William Shakespeare](https://www.gutenberg.org/ebooks/100) with the goal of the next character prediction.
19
+
20
+ ## Dataset Details
21
+
22
+ ### Dataset Description
23
+
24
+ Each sample is comprised of a text of 80 characters (x) and a next character (y).
25
+
26
+ - **Curated by:** [LEAF](https://leaf.cmu.edu/)
27
+ - **Language(s) (NLP):** English
28
+ - **License:** BSD 2-Clause License
29
+
30
+ ### Dataset Sources
31
+
32
+ The code from the original repository was adopted to post it here.
33
+
34
+ - **Repository:** https://github.com/TalwalkarLab/leaf
35
+ - **Paper:** https://arxiv.org/abs/1812.01097
36
+
37
+ ## Uses
38
+
39
+ This dataset is intended to be used in Federated Learning settings.
40
+ A pair of a character and a play denotes a unique user in the federation.
41
+
42
+ ### Direct Use
43
+
44
+ <!-- This section describes suitable use cases for the dataset. -->
45
+
46
+ [More Information Needed]
47
+
48
+ ### Out-of-Scope Use
49
+
50
+ <!-- This section addresses misuse, malicious use, and uses that the dataset will not work well for. -->
51
+
52
+ [More Information Needed]
53
+
54
+ ## Dataset Structure
55
+
56
+ The dataset contains only train split. The split in the paper happens at each node only (no centralized dataset).
57
+ The dataset is comprised of columns:
58
+ * `character_id` - id denoting a pair of character + play (node in federated learning settings)
59
+ * `x`: str - text of 80 characters
60
+ * `y`: str - single character following the `x`
61
+ Please note that the data is temporal. Therefore, caution is needed when dividing it so as not to leak the information from the train set.
62
+
63
+ [More Information Needed]
64
+
65
+ ## Dataset Creation
66
+
67
+ ### Curation Rationale
68
+
69
+ This dataset was created as a part of the [LEAF](https://leaf.cmu.edu/) benchmark.
70
+
71
+ ### Source Data
72
+
73
+ [The Complete Works of William Shakespeare](https://www.gutenberg.org/ebooks/100)
74
+
75
+ #### Data Collection and Processing
76
+
77
+ For the preprocessing details, please refer to the original paper and the source code.
78
+
79
+ #### Who are the source data producers?
80
+
81
+ William Shakespeare
82
+
83
+ ## Citation
84
+
85
+ When working on the LEAF benchmark, please cite the original paper. If you're using this dataset with Flower Datasets, you can cite Flower.
86
+
87
+ **BibTeX:**
88
+ ```
89
+ @article{DBLP:journals/corr/abs-1812-01097,
90
+ author = {Sebastian Caldas and
91
+ Peter Wu and
92
+ Tian Li and
93
+ Jakub Kone{\v{c}}n{\'y} and
94
+ H. Brendan McMahan and
95
+ Virginia Smith and
96
+ Ameet Talwalkar},
97
+ title = {{LEAF:} {A} Benchmark for Federated Settings},
98
+ journal = {CoRR},
99
+ volume = {abs/1812.01097},
100
+ year = {2018},
101
+ url = {http://arxiv.org/abs/1812.01097},
102
+ eprinttype = {arXiv},
103
+ eprint = {1812.01097},
104
+ timestamp = {Wed, 23 Dec 2020 09:35:18 +0100},
105
+ biburl = {https://dblp.org/rec/journals/corr/abs-1812-01097.bib},
106
+ bibsource = {dblp computer science bibliography, https://dblp.org}
107
+ }
108
+ ```
109
+ ```
110
+ @article{DBLP:journals/corr/abs-2007-14390,
111
+ author = {Daniel J. Beutel and
112
+ Taner Topal and
113
+ Akhil Mathur and
114
+ Xinchi Qiu and
115
+ Titouan Parcollet and
116
+ Nicholas D. Lane},
117
+ title = {Flower: {A} Friendly Federated Learning Research Framework},
118
+ journal = {CoRR},
119
+ volume = {abs/2007.14390},
120
+ year = {2020},
121
+ url = {https://arxiv.org/abs/2007.14390},
122
+ eprinttype = {arXiv},
123
+ eprint = {2007.14390},
124
+ timestamp = {Mon, 03 Aug 2020 14:32:13 +0200},
125
+ biburl = {https://dblp.org/rec/journals/corr/abs-2007-14390.bib},
126
+ bibsource = {dblp computer science bibliography, https://dblp.org}
127
+ }
128
+ ```
129
+
130
+ ## Dataset Card Contact
131
+
132
+ In case of any doubts, please contact Flower Labs.