language:
- en
- fr
license: mit
task_categories:
- text-generation
- conversational
- text2text-generation
tags:
- chat
- code
- reasoning
- motif
- mega-dataset
size_categories:
- 100M<n<1B
---

# Motif Mega Chat Code

## Description
A massive dataset for training Motif with chat + code capabilities.

## Content
- **25,000,000+** examples
- **8 categories**: code requests, code explanations, debugging, optimization, documentation, tests, architecture, theory
- **Languages**: English, French

## Categories

| Category | % | Examples |
|----------|---|----------|
| Code requests | 20% | 5,000,000 |
| Code explanations | 15% | 3,750,000 |
| Debugging | 15% | 3,750,000 |
| Optimization | 10% | 2,500,000 |
| Documentation | 10% | 2,500,000 |
| Tests | 10% | 2,500,000 |
| Architecture | 10% | 2,500,000 |
| Theory | 10% | 2,500,000 |

## Usage

```python
from datasets import load_dataset

dataset = load_dataset("Lelonthecodeur/code-and-chat-dataset")