el2e10/aya-indicsentiment-malayalam

---
language:
- ml
license: cc
size_categories:
- 1K<n<10K
task_categories:
- conversational
dataset_info:
features:
- name: inputs
dtype: string
- name: targets
dtype: string
- name: template_lang
dtype: string
- name: template_id
dtype: int64
splits:
- name: train
num_bytes: 725805
num_examples: 1156
download_size: 315813
dataset_size: 725805
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---
### Description

This dataset is derived from the already existing dataset made by AI4Bharat. We have used the [IndicSentiment](https://huggingface.co/datasets/ai4bharat/IndicSentiment) dataset of AI4Bharat to create an instruction style dataset.

IndicSentiment is a multilingual parallel dataset for sentiment analysis. It encompasses product reviews, translations into Indic languages, sentiment labels, and more.
The original dataset(IndicSentiment) was made available under the cc-0 license.

This dataset comprises a singular split (train) and encompasses over 1150+ rows.

### Template

The following template was used for converting the original dataset:

```
#Template 1
prompt:
Translate from English to {target_language}:
{ENGLISH_REVIW}

completion:
{INDIC_REVIEW}
```
```
#Template 2
prompt:
Translate this sentence to {target_language}:
{ENGLISH_REVIW}

completion:
{INDIC_REVIEW}
```
```
#Template 3
prompt:
What's the {target_language} translation of this language:
{ENGLISH_REVIW}

completion:
{INDIC_REVIEW}
```
```
#Template 4
prompt:
Can you translate this text to {target_language}:
{ENGLISH_REVIW}

completion:
{INDIC_REVIEW}
```

原始信息汇总

数据集概述

基本信息

• 语言: 马拉雅拉姆语 (ml)
• 许可证: Creative Commons (cc)
• 数据规模: 1K<n<10K
• 任务类别: 对话 (conversational)

数据集结构

• 特征:

  • inputs: 字符串类型
  • targets: 字符串类型
  • template_lang: 字符串类型
  • template_id: 64位整数类型

• 分割:

  • train:
    • 字节数: 725805
    • 样本数: 1156

• 下载大小: 315813

• 数据集大小: 725805

配置

• 默认配置:
  • 数据文件:
    • 分割: train
    • 路径: data/train-*

描述

该数据集源自AI4Bharat已有的IndicSentiment数据集，用于创建指令样式数据集。IndicSentiment是一个用于情感分析的多语言平行数据集，包含产品评论、印度语言翻译、情感标签等。原始数据集(IndicSentiment)在cc-0许可证下发布。

该数据集包含一个训练分割(train)，包含超过1150行数据。

模板

以下模板用于转换原始数据集:

#Template 1 prompt: Translate from English to {target_language}: {ENGLISH_REVIW}

completion: {INDIC_REVIEW}

#Template 2 prompt: Translate this sentence to {target_language}: {ENGLISH_REVIW}

completion: {INDIC_REVIEW}

#Template 3 prompt: Whats the {target_language} translation of this language: {ENGLISH_REVIW}

completion: {INDIC_REVIEW}

#Template 4 prompt: Can you translate this text to {target_language}: {ENGLISH_REVIW}

completion: {INDIC_REVIEW}