Dataset · bigcode
GitHub Issues and Pull Requests discussions (54 GB)
bigcode/the-stack-github-issues
A corpus of 30.9 million files featuring developer discussions in GitHub issues and pull requests.
What's inside
A large 54 GB corpus from the BigCode project containing 30.9 million files with event chains from GitHub Issues and Pull Requests, predominantly in English. Each branch includes an action sequence (opening, comments, closing), message texts, author logins, and task identifiers.
What you can do with it
- Training code LLMs to conduct dialogues on bug reports and pull request reviews
- Creating AI agents for automatic triage and classification of tasks in repositories
- Tasks
- Text generation
- Languages
- English
- Size
- 10M–100M rows
- Downloads this month
- 492
- Likes
- 53
- Access
- on request on Hugging Face
Models trained on it
- stabilityai/stable-code-3bText generation 6.1K
- TheBloke/stable-code-3b-GGUF 2.1K
- mradermacher/stable-code-3b-i1-GGUF 2K
- mradermacher/stable-code-3b-GGUF 1.7K
- QuantFactory/stable-code-3b-GGUFText generation 1.2K
- brittlewis12/stable-code-3b-GGUFText generation 456
- TheBloke/stable-code-3b-GPTQText generation 215
- DhruvalLabs/stable-code-3b-GGUFText generation 211
How to load it
from datasets import load_dataset
ds = load_dataset("bigcode/the-stack-github-issues")