Dataset · HuggingFaceH4
CodeAlpaca 20K: train and test split
HuggingFaceH4/CodeAlpaca_20K
Original CodeAlpaca split into train and test; about 20k text examples for code generation.
What's inside
The dataset represents the original CodeAlpaca split into training and test samples. The size is between 10k and 100k examples, and the modality is text. The language and detailed composition are not specified in the card.
What you can do with it
- Training code generation models
- Evaluating on the CodeAlpaca test split
- SFT for a code assistant
- Tasks
- Text generation
- Size
- 10K–100K rows
- Downloads this month
- 5.1K
- Likes
- 116
Models trained on it
- askdkc/DS-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-MetalText generation 2.8K
- TheBloke/Llama-2-Coder-7B-GGUFText generation 2.7K
- mradermacher/Dolphy-1.0-i1-GGUF 750
- mradermacher/Bootes-Qwen3_Coder-Reasoning-i1-GGUF 686
- mradermacher/code-gemma-2b-it-GGUF 597
- mradermacher/Dolphy-1.0-GGUF 526
- hsam007/Nanbeige4.3-33B-NVFP4Text generation 423
- mradermacher/Bootes-Qwen3_Coder-Reasoning-GGUF 393
How to load it
from datasets import load_dataset
ds = load_dataset("HuggingFaceH4/CodeAlpaca_20K")