to make it fair comparison a bit 😄 All benchmarks' results are an average of 100 runs. See our benchmarks.py for methodology and to run your comparisons. This library is provided for educational and ...