Skip to content

Dataset · theblackcat102

English GPT-4 coding instructions

theblackcat102/evol-codealpaca-v1

Programming instruction-response pairs augmented by GPT-4 for training code assistants.

What's inside

The dataset contains English programming instructions and responses generated and augmented by GPT-4 models, mainly gpt-4-0314. The methodology is close to WizardCoder, but the set is open. After filtering, only clean English instructions remain without mentions of OAI-created responses. Median sequence length is 471 tokens, volume ranges from 100k to 1m examples.

What you can do with it

  • SFT of code assistants on English instructions
  • Comparison with WizardCoder-like sets
  • Evaluating code generation from instructions
Tasks
Text generation
Languages
English
Size
100K–1M rows
Downloads this month
9.3K
Likes
194

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("theblackcat102/evol-codealpaca-v1")