Dataset · Nan-Do
Python code generation and description instructions
Nan-Do/instructional_code-search-net-python
English instructional dataset for generating Python code from descriptions and explaining ready code.
What's inside
The set contains between 100,000 and 1,000,000 English text instructions for working with Python code. It features two types of tasks: composing a description for a code snippet and writing working code based on a given description. The dataset was collected in May 2023 and is not split.
What you can do with it
- Fine-tuning coding LLMs and LoRA for Python function generation using text specifications
- Training models for automatic documentation and explanation of external code
- Tasks
- Text generation
- Languages
- English
- Size
- 100K–1M rows
- Downloads this month
- 520
- Likes
- 38
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("Nan-Do/instructional_code-search-net-python")