Update README.md
Browse files
README.md
CHANGED
@@ -14,4 +14,83 @@ language_creators:
|
|
14 |
multilinguality:
|
15 |
- monolingual
|
16 |
pretty_name: Thirukkural_QA
|
17 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
14 |
multilinguality:
|
15 |
- monolingual
|
16 |
pretty_name: Thirukkural_QA
|
17 |
+
---
|
18 |
+
|
19 |
+
# Summary
|
20 |
+
`thirukkural_QA` is an open source dataset of instruct-style records generated by converting publicly available data on Thirukkural and it's meaning.
|
21 |
+
This was created as part of [Aya Open Science Initiative](https://sites.google.com/cohere.com/aya-en/home) by Cohere For AI.
|
22 |
+
|
23 |
+
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.
|
24 |
+
|
25 |
+
Supported Tasks:
|
26 |
+
- Training LLMs
|
27 |
+
- Synthetic Data Generation
|
28 |
+
- Data Augmentation
|
29 |
+
- Question Answering
|
30 |
+
|
31 |
+
Languages: Tamil Version: 1.0
|
32 |
+
|
33 |
+
# Dataset Overview
|
34 |
+
`thirukkural_QA` is a corpus of 3990 records generated by converting existing Thirukkural and its meaning into instruction-style. This Dataset can be used for the following tasks:
|
35 |
+
- Given the thirukkural and ask for its meaning, generates the meaning of the kural.
|
36 |
+
- Given the meaning of the kural, generates the original kural.
|
37 |
+
- Given the beginning of a kural, generates the original kural.
|
38 |
+
|
39 |
+
# Intended Uses
|
40 |
+
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. For example, prompt-completions could be submitted as few-shot examples to a large open language model to generate new kurals in a similar style.
|
41 |
+
|
42 |
+
# Dataset
|
43 |
+
|
44 |
+
## Load with Datasets
|
45 |
+
To load this dataset with Datasets, you'll just need to install Datasets as `pip install datasets --upgrade` and then use the following code:
|
46 |
+
|
47 |
+
```python
|
48 |
+
from datasets import load_dataset
|
49 |
+
ds = load_dataset('aitamilnadu/thirukkural_QA')
|
50 |
+
```
|
51 |
+
|
52 |
+
## Purpose of Collection
|
53 |
+
Tamil is a low-resource language (inspite of having rich literature) where there are no instruct-style dataset to the best of my knowledge.
|
54 |
+
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.
|
55 |
+
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.
|
56 |
+
|
57 |
+
## Sources
|
58 |
+
- **[Thirukkural.com](http://www.thirukkural.com/)**: The data from this website is scraped and available at [Thirukkural-Tamil-Dataset](https://github.com/vijayanandrp/Thirukkural-Tamil-Dataset).
|
59 |
+
- The scraped data is carefully analysed making sure there are no missed words, spelling mistakes and the data is in Tamil only.
|
60 |
+
- Next, some pre-processing is performed to extract kural, adhigaram, kural no and different meanings seperately from the scraped data.
|
61 |
+
- Finally, converted the scraped data into instruct-style prompts and completions.
|
62 |
+
|
63 |
+
|
64 |
+
## Templates
|
65 |
+
For the creation of instruct-style prompts and completions from the scraped data, the following one template was used:
|
66 |
+
1. Given the name of the food item, generates the detailed recipe along with the ingredients.
|
67 |
+
```python
|
68 |
+
Prompt:
|
69 |
+
("{{Food Item}} ఎలా తయారు చేస్తారో క్లుప్తంగా ఇవ్వండి." /
|
70 |
+
"మీరు {{Food Item}} తయారు చెయ్యడానికి ఎటువంటి తిండి పదార్ధాలు వాడుతురు మరియు ఏ విధముగా చేస్తారో వివరణ ఇవ్వండి." /
|
71 |
+
"నేను {{Food Item}} చాలా రుచికరంగా ఉంటుంది అని విన్నాను. నాకు ఇది ఎలా చెయ్యాలో సంక్లిప్తంగా చెప్పు." /
|
72 |
+
"మొదటిసారి వంట చేసేవారికి చెప్పినట్టు నాకు {{Food Item}} ఎలా చెయ్యాలొ చెప్పు" /
|
73 |
+
"{{Food Item}} ఎలా చేయాలి? సమాధానం లో కావలసిన పదార్ధాలు మరియు తయారు చేసే విధానం ఉండాలి." /
|
74 |
+
"{{Food Item}} ఎలా తయారు చేస్తాం?" /
|
75 |
+
"{{Food Item}} రెసిపీ ఏంటి?")
|
76 |
+
|
77 |
+
Completion:
|
78 |
+
{{Food Item}} కి కావలసిన పదార్ధాలు మరియు తయారు చేసే విధానం ఇక్కడ ఇవ్వబడింది.
|
79 |
+
కావలసిన పదార్థాలు:
|
80 |
+
{{Ingredients}}
|
81 |
+
|
82 |
+
తయారుచేసే విధానం:
|
83 |
+
{{Cooking Instructions}}
|
84 |
+
```
|
85 |
+
|
86 |
+
## Personal or Sensitive Data
|
87 |
+
This dataset contains public information. To my knowledge, there are no private person’s personal identifiers or sensitive information.
|
88 |
+
|
89 |
+
## Language
|
90 |
+
Tamil
|
91 |
+
|
92 |
+
# Known Limitations
|
93 |
+
- 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.
|
94 |
+
|
95 |
+
# Contributors
|
96 |
+
[AbinayaM02](https://github.com/AbinayaM02)
|