Skip to content

khulnasoft/eval

 
 

Repository files navigation

HumanEval for GPT-3.5/GPT-4

To generate the completions (after pip installing the requirements), run:

mkdir results
python run.py

Then to evaluate the completion results, run

pip3 install -e .
evaluate_functional_correctness results/name_of_results_file

About

Code for the paper "Evaluating Large Language Models Trained on Code"

Resources

License

Code of conduct

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Python 100.0%