emirkaanozdemr/bash_command_data_6K

---
dataset_info:
features:
- name: prompt
dtype: string
- name: completion
dtype: string
splits:
- name: train
num_bytes: 942238
num_examples: 6153
download_size: 476358
dataset_size: 942238
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: apache-2.0
task_categories:
- text-generation
language:
- en
tags:
- code
pretty_name: lds
size_categories:
- 1K<n<10K
---

# 📦 Bash Command Dataset v1

A high-quality dataset of **natural language instructions paired with their equivalent Bash commands**, designed for training and fine-tuning large language models (LLMs) that translate English tasks into shell commands.

This dataset is ideal for researchers, developers, and machine learning engineers interested in **natural language to Bash command translation**, command-line automation, and building intelligent terminal assistants.

---

## 📁 Dataset Structure

The dataset contains a single split (`train`) in JSONL / JSON format. Each example consists of two fields:

- **`prompt`** *(string)*: A natural language description of a task that a user wants to perform in a Unix-like shell.
- **`completion`** *(string)*: The corresponding Bash command (or command sequence) that fulfills the described task.

### Example Entry

```json
{"prompt": "List all files in the current directory, including hidden ones, in long format.", "completion": "ls -la\n"}
```
# 🔧 Use Cases

This dataset can be used for:

- 🧠 **Fine-tuning LLMs** to convert English instructions into Bash commands.
- 💻 **Building AI assistants** for command-line automation.
- 📊 **Evaluating model performance** on shell command generation tasks.
- 🚀 **Research** on NL2SH (Natural Language to Shell) translation systems.

## ⚠️ Safety and Usage Notes

- **Execution Caution:** Some commands may be destructive (e.g., deletion of files). Always execute generated commands in a safe environment (sandbox or container) before running them on real systems.
- **Environment Specifics:** Commands are written for Linux-like systems (e.g., Ubuntu). Some commands or options might behave differently on other shells or distributions.
- Responsible use is strongly recommended.

## 📜 License

This dataset is shared under the **Apache 2.0 License**

## 📈 Download and Usage

You can load the dataset directly using the Hugging Face Datasets library:

```python
from datasets import load_dataset

dataset = load_dataset("emirkaanozdemr/bash_command_data_6K")
```
# 📚 Citation

If you use this dataset in your work, please cite it as:
```latex
@misc{ozdemir2026bash,
author = {Emir Kaan Ozdemir},
title = {Bash Command Dataset 6K},
year = {2026},
publisher = {Hugging Face},
journal = {Hugging Face Dataset},
url = {https://huggingface.co/datasets/emirkaanozdemr/bash_command_data_6K}
}