sartifyllc/clean_loal_pretrain

---
language:
- sw
tags:
- dataset
- swahili
license: mit
---

# Dataset Card for sartifyllc/clean_loal_pretrain

## Dataset Description

This is a Swahili dataset for CPT.

## Dataset Usage

This dataset can be used for [describe potential uses].

## Dataset Structure

### Features:

- text: Number of rows 3017
- token: Number of token so far 0.003997569B
- token per row: Number of token so far per row 1325.0145840238647

## Dataset Creation

### Source Data

- Homepage: https://huggingface.co/datasets/sartifyllc/clean_loal_pretrain

### Data Preprocessing

[Describe any preprocessing steps here]

## Considerations for Using the Data

[Add any considerations, such as ethical concerns or potential biases]

## Additional Information

[Any other relevant information about the dataset]