It's mentioned in neither the codebase nor the accompanied paper, so I wonder if you recall the performance improvements from the cythonization, vs interpreting CPython bytecode and PyPy's JIT optimizations.
I'm trying to justify the Cython compilation overhead while hacking on the DuoGlot server.
It's mentioned in neither the codebase nor the accompanied paper, so I wonder if you recall the performance improvements from the cythonization, vs interpreting CPython bytecode and PyPy's JIT optimizations.
I'm trying to justify the Cython compilation overhead while hacking on the DuoGlot server.