---
tags:
- code
- programming
- dataset
pretty_name: "Coding Dataset"
---

# Coding Dataset

Production-grade dataset for training AI coding agents.

## Dataset Summary

- **Total Examples**: 6 (demo)
- **Languages**: Python, JavaScript, Java
- **Task Types**: Code Generation
- **License**: CC0-1.0

## Dataset Structure

### Data Splits

- train: 70% of data
- validation: 15% of data  
- test: 15% of data

### Features

- `id` (string): Unique identifier
- `code` (string): Source code snippet
- `code_description` (string): Natural language description
- `programming_language` (string): Language (python, javascript, java, etc.)
- `task_type` (string): Type of task
- `difficulty_level` (string): Difficulty (beginner, intermediate, advanced, expert)
- `quality_score` (float): Quality score 0.0-1.0
- `is_tested` (bool): Code is tested
- `has_bugs` (bool): Known bugs exist
- `lines_of_code` (int): Number of lines
- `collected_at` (string): Collection timestamp

## Usage

```python
from datasets import load_dataset

# Load dataset
dataset = load_dataset("romcmu863/code-dataset")

# Access splits
train = dataset['train']
validation = dataset['validation']
test = dataset['test']

# Get first example
example = train[0]
print(example['code_description'])
print(example['code'])
```

## License

CC0-1.0

## Created

2025-10-25