README.md

---
language:
- as
- bn
- brx
- doi
- en
- gom
- gu
- hi
- kn
- ks
- mai
- ml
- mr
- mni
- ne
- or
- pa
- sa
- sat
- sd
- ta
- te
- ur
language_details: >-
  asm_Beng, ben_Beng, brx_Deva, doi_Deva, eng_Latn, gom_Deva, guj_Gujr,
  hin_Deva, kan_Knda, kas_Arab, mai_Deva, mal_Mlym, mar_Deva, mni_Mtei, 
  npi_Deva, ory_Orya, pan_Guru, san_Deva, sat_Olck, snd_Deva, tam_Taml, 
  tel_Telu, urd_Arab
license: cc-by-4.0
language_creators:
- expert-generated
multilinguality:
- multilingual
- translation
pretty_name: in22-conv
size_categories:
- 1K<n<10K
task_categories:
- translation
---

# IN22-Conv

IN-22 is a newly created comprehensive benchmark for evaluating machine translation performance in multi-domain, n-way parallel contexts across 22 Indic languages. IN22-Conv is the conversation domain subset of IN22. It is designed to assess translation quality in typical day-to-day conversational-style applications. The evaluation subset consists of 1503 sentences translated across 22 Indic languages enabling evaluation of MT systems across 506 directions.

Currently, we use it for sentence-level evaluation of MT systems but it can be repurposed for document translation evaluation as well.

Here is the domain distribution of our IN22-Conv evaluation subset.

<table style="width:25%">
    <tr>
        <td>domain</td>
        <td>count</td>
    </tr>
    <tr>
        <td>hobbies</td>
        <td>120</td>
    </tr>
    <tr>
        <td>daily_dialogue</td>
        <td>117</td>
    </tr>
    <tr>
        <td>government</td>
        <td>116</td>
    </tr>
    <tr>
        <td>geography</td>
        <td>114</td>
    </tr>
    <tr>
        <td>sports</td>
        <td>100</td>
    </tr>
    <tr>
        <td>entertainment</td>
        <td>97</td>
    </tr>
    <tr>
        <td>history</td>
        <td>97</td>
    </tr>
    <tr>
        <td>legal</td>
        <td>96</td>
    </tr>
    <tr>
        <td>arts</td>
        <td>95</td>
    </tr>
    <tr>
        <td>college_life</td>
        <td>94</td>
    </tr>
    <tr>
        <td>tourism</td>
        <td>91</td>
    </tr>
    <tr>
        <td>school_life</td>
        <td>87</td>
    </tr>
    <tr>
        <td>insurance</td>
        <td>82</td>
    </tr>
    <tr>
        <td>culture</td>
        <td>73</td>
    </tr>
    <tr>
        <td>healthcare</td>
        <td>67</td>
    </tr>
    <tr>
        <td>banking</td>
        <td>57</td>
    </tr>
    <tr>
        <td>total</td>
        <td>1503</td>
    </tr>
</table>

Please refer to the `Appendix E: Dataset Card` of the [preprint](https://arxiv.org/abs/2305.16307) on detailed description of dataset curation, annotation and quality control process.


### Dataset Structure

#### Dataset Fields

- `id`: Row number for the data entry, starting at 1.
- `doc_id`: Unique identifier of the conversation.
- `sent_id`: Unique identifier of the sentence order in each conversation.
- `topic`: The specific topic of the conversation within the domain.
- `domain`: The domain of the conversation.
- `prompt`: The prompt provided to annotators to simulate the conversation.
- `scenario`: The scenario or context in which the conversation takes place.
- `speaker`: The speaker identifier in the conversation.
- `turn`: The turn within the conversation.

#### Data Instances

A sample from the `gen` split for the English language (`eng_Latn` config) is provided below. All configurations have the same structure, and all sentences are aligned across configurations and splits.

```python
{
   "id": 1,
   "doc_id": 0,
   "sent_id": 1,
   "topic": "Festivities",
   "domain": "culture",
   "prompt": "14th April a holiday",
   "scenario": "Historical importance",
   "speaker": 1,
   "turn": 1,
   "sentence": "Mom, let's go for a movie tomorrow."
}
```

When using a hyphenated pairing or using the `all` function, data will be presented as follows:

```python
{
   "id": 1,
   "doc_id": 0,
   "sent_id": 1,
   "topic": "Festivities",
   "domain": "culture",
   "prompt": "14th April a holiday",
   "scenario": "Historical importance",
   "speaker": 1,
   "turn": 1,
   "sentence_eng_Latn": "Mom, let's go for a movie tomorrow.",
   "sentence_hin_Deva": "माँ, चलो कल एक फिल्म देखने चलते हैं।"
}
```

#### Sample Conversation

<table>
    <tr>
        <td>Speaker</td>
        <td>Turn</td>
    </tr>
    <tr>
        <td>Speaker 1</td>
        <td>Mom, let&#39;s go for a movie tomorrow. I don&#39;t have to go to school. It is a holiday.</td>
    </tr>
    <tr>
        <td>Speaker 2</td>
        <td>Oh, tomorrow is the 14th of April right? Your dad will also have the day off from work. We can make a movie plan!</td>
    </tr>
    <tr>
        <td>Speaker 1</td>
        <td>That&#39;s a good news! Why is it a holiday though? Are all schools, colleges and offices closed tomorrow?</td>
    </tr>
    <tr>
        <td>Speaker 2</td>
        <td>It is Ambedkar Jayanti tomorrow! This day is celebrated annually to mark the birth of Dr. B. R Ambedkar. Have you heard of him?</td>
    </tr>
    <tr>
        <td>Speaker 1</td>
        <td>I think I have seen him in my History and Civics book. Is he related to our Constitution?</td>
    </tr>
    <tr>
        <td>Speaker 2</td>
        <td>Absolutely! He is known as the father of the Indian Constitution. He was a civil rights activist who played a major role in formulating the Constitution. He played a crucial part in shaping the vibrant democratic structure that India prides itself upon.</td>
    </tr>
    <tr>
        <td></td>
        <td>...</td>
    </tr>
</table>


### Usage Instructions

```python
from datasets import load_dataset

# download and load all the pairs
dataset = load_dataset("ai4bharat/IN22-Conv", "all")

# download and load specific pairs
dataset = load_dataset("ai4bharat/IN22-Conv", "eng_Latn-hin_Deva")
```

### Languages Covered

<table style="width: 40%">
    <tr>
        <td>Assamese (asm_Beng)</td>
        <td>Kashmiri (Arabic) (kas_Arab)</td>
        <td>Punjabi (pan_Guru)</td>
    </tr>
    <tr>
        <td>Bengali (ben_Beng)</td>
        <td>Kashmiri (Devanagari) (kas_Deva)</td>
        <td>Sanskrit (san_Deva)</td>
    </tr>
    <tr>
        <td>Bodo (brx_Deva)</td>
        <td>Maithili (mai_Deva)</td>
        <td>Santali (sat_Olck)</td>
    </tr>
    <tr>
        <td>Dogri (doi_Deva)</td>
        <td>Malayalam (mal_Mlym)</td>
        <td>Sindhi (Arabic) (snd_Arab)</td>
    </tr>
    <tr>
        <td>English (eng_Latn)</td>
        <td>Marathi (mar_Deva)</td>
        <td>Sindhi (Devanagari) (snd_Deva)</td>
    </tr>
    <tr>
        <td>Konkani (gom_Deva)</td>
        <td>Manipuri (Bengali) (mni_Beng)</td>
        <td>Tamil (tam_Taml)</td>
    </tr>
    <tr>
        <td>Gujarati (guj_Gujr)</td>
        <td>Manipuri (Meitei) (mni_Mtei)</td>
        <td>Telugu (tel_Telu)</td>
    </tr>
    <tr>
        <td>Hindi (hin_Deva)</td>
        <td>Nepali (npi_Deva)</td>
        <td>Urdu (urd_Arab)</td>
    </tr>
    <tr>
        <td>Kannada (kan_Knda)</td>
        <td>Odia (ory_Orya)</td>
    </tr>
</table>


### Citation

If you consider using our work then please cite using:

```
@article{gala2023indictrans,
title={IndicTrans2: Towards High-Quality and Accessible Machine Translation Models for all 22 Scheduled Indian Languages},
author={Jay Gala and Pranjal A Chitale and A K Raghavan and Varun Gumma and Sumanth Doddapaneni and Aswanth Kumar M and Janki Atul Nawale and Anupama Sujatha and Ratish Puduppully and Vivek Raghavan and Pratyush Kumar and Mitesh M Khapra and Raj Dabre and Anoop Kunchukuttan},
journal={Transactions on Machine Learning Research},
issn={2835-8856},
year={2023},
url={https://openreview.net/forum?id=vfT4YuzAYA},
note={}
}
```