abinayam commited on
Commit
6814ce3
1 Parent(s): 98f8ae6

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +98 -0
README.md CHANGED
@@ -16,4 +16,102 @@ multilinguality:
16
  pretty_name: tamil_stories
17
  ---
18
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
19
 
 
16
  pretty_name: tamil_stories
17
  ---
18
 
19
+ # Summary
20
+ `tamil_stories` is an open source dataset of instruct-style records generated by scraping publicly available short stories on the following websites.
21
+ - [Siruvarmalar](https://www.siruvarmalar.com/)
22
+ - [Tamilsurangam](https://www.tamilsurangam.in/)
23
+
24
+ Apart from scraping and automated cleaning, the data was also tagged manually by a group of volunteers.
25
+ This dataset created as part of [Aya Open Science Initiative](https://sites.google.com/cohere.com/aya-en/home) by Cohere For AI.
26
+ This dataset can be used for any purpose, whether academic or commercial, under the terms of the [Apache 2.0](https://opensource.org/license/apache-2-0) License.
27
+
28
+ Supported Tasks:
29
+ - Training LLMs
30
+ - Synthetic Data Generation
31
+ - Data Augmentation
32
+ - Question Answering
33
+
34
+ Languages: Tamil Version: 1.0
35
+
36
+ # Dataset Overview
37
+ `tamil_stories` is a corpus of 1202 records generated by converting scraped stories into instruction-style. This Dataset can be used for the following tasks:
38
+ - Given the story, generate the appropriate title for the story.
39
+ - Given some prominent characters/words from the story along with the title, generate the complete story.
40
+
41
+
42
+ # Intended Uses
43
+ While immediately valuable for instruction fine tuning large language models, as a corpus of instruction prompts, this dataset also presents a valuable opportunity for synthetic data generation.
44
+ For example, prompt-completions could be submitted as few-shot examples to a large open language model to generate new stories in a similar style.
45
+
46
+ # Dataset
47
+
48
+ ## Load with Datasets
49
+ To load this dataset with Datasets, you'll just need to install Datasets as `pip install datasets --upgrade` and then use the following code:
50
+
51
+ ```python
52
+ from datasets import load_dataset
53
+ ds = load_dataset('aitamilnadu/tamil_stories')
54
+ ```
55
+
56
+ ## Purpose of Collection
57
+ Tamil is a low-resource language (inspite of having rich literature) where there are no instruct-style dataset to the best of my knowledge.
58
+ This was created as a part of [Aya Open Science Initiative](https://sites.google.com/cohere.com/aya-en/home) from Cohere For AI to make sure Tamil is well represented in the space of AI/ML.
59
+ Unlike other datasets that are limited to non-commercial use, this dataset can be used, modified, and extended for any purpose, including academic or commercial applications.
60
+
61
+ ## Sources
62
+ The following are the websites from which the data is scraped.
63
+ - [Siruvarmalar](https://www.siruvarmalar.com/)
64
+ - [Tamilsurangam](https://www.tamilsurangam.in/)
65
+
66
+ The scripts used for scraping and cleaning the data can be found here:
67
+ [Tamilsurangam Webscraper](https://github.com/miluckshan-j/tamilsurangam-webscraper),
68
+ [Siruvarmalar Webscraper](https://github.com/miluckshan-j/siruvarmalar-webscraper)
69
+ [Dataset Creation](https://github.com/miluckshan-j/tamil-dataset-creation-tools)
70
+
71
+ - The scraped data is carefully analysed making sure there are no missed words, spelling mistakes and the data is in Tamil only.
72
+ - Words/characters are manually added to stories by volunteers.
73
+ - Finally, converted the pre-processed data into instruct-style prompts and completions.
74
+
75
+
76
+ ## Templates
77
+ For the creation of instruct-style prompts and completions from the scraped data, the following templates were used:
78
+ 1. Given the story, generate the appropriate title for the story.
79
+
80
+ ```python
81
+ Prompt:
82
+ கீழே கொடுக்கப்பட்டுள்ள கதைக்குப் பொருத்தமான தலைப்பைக் கொடு.
83
+ கதை: {Story}
84
+
85
+ Completion:
86
+ கொடுக்கப்பட்டுள்ள கதைக்குப் பொருத்தமான தலைப்பு '{Title}' என்பதாகும்.
87
+ ```
88
+
89
+ 2. Given some prominent characters/words from the story along with the title, generate the complete story.
90
+ ```python
91
+ Prompt:
92
+ கீழே கொடுக்கப்பட்டுள்ள வார்த்தைகளையும் தலைப்பையும் பயன்படுத்தி சிறு கதை எழுதுக.
93
+ வார்த்தைகள்: {Comma_Seperated_Words}
94
+ தலைப்பு: {Title}
95
+
96
+ Completion:
97
+ {Story}
98
+ ```
99
+
100
+ ## Personal or Sensitive Data
101
+ This dataset contains public information. To my knowledge, there are no private person’s personal identifiers or sensitive information.
102
+
103
+ ## Language
104
+ Tamil
105
+
106
+ # Known Limitations
107
+ - The meanings used in the prompts/completions are chosen randomly based on the availability of complete sentences and this may reflect some bias by ignoring other meanings written by other scholars.
108
+
109
+ # Contributors
110
+ [miluckshan-j](https://github.com/miluckshan-j)
111
+ [chezhian0599](https://github.com/chezhian0599)
112
+ [AbinayaM02](https://github.com/AbinayaM02)
113
+ [anitha67](https://github.com/anitha67)
114
+ [vijiv11](https://github.com/vijiv11)
115
+ [Lagstill](https://github.com/Lagstill)
116
+ [Merlinecodes](https://github.com/Merlinecodes)
117