README.md

---
license: apache-2.0
language:
- sw
- en
tags:
- kenyan
- conversational
- sheng
- swahili
- instruction-tuning
dataset_info:
  features:
  - name: category
    dtype: string
  - name: language_style
    dtype: string
  - name: user
    dtype: string
  - name: assistant
    dtype: string
  splits:
  - name: train
    num_bytes: 543430
    num_examples: 15552
  download_size: 8351
  dataset_size: 543430
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

# Kenyan Conversational Dataset

Hii ni dataset ya mazungumzo iliyoundwa kwa ajili ya kufunza AI kuelewa lugha ya mtaani ya Kenya (Sheng, Swahili, na Kenyan English mix).

## Data Details

Dataset hii ina mifano ya mazungumzo ya kila siku, ikijumuisha:
* Salamu za kawaida (Greetings)
* Masuala ya Uchumi na Siasa
* Maisha ya mtaani (Matatu, Boda, Hustle, Payday)
* Teknolojia na Simu

## Data Structure
Kila mfano (example) una muundo ufuatao:
* `category`: Aina ya mazungumzo.
* `language_style`: Mtindo wa lugha (Sheng, WhatsApp style, n.k.).
* `user`: Ujumbe kutoka kwa mtumiaji.
* `assistant`: Jibu kutoka kwa AI.

## Usage
Unaweza kutumia dataset hii kama ifuatavyo:
```python
from datasets import load_dataset
dataset = load_dataset("micymike/kenyan-conversational-data")
```