statsmodels: QuantReg runs 30x slower than quantile regression in R

Hi Guys, I face a challenge as following:

Response variables from a pandas DataFrame: endog, 5000 * 90 Explanatory variable: exog, a pandas Series, length = 90 ---- session info — OS: macos RAM: 16G CPU: core i7 ---------=----------- from statsmodels.tools import tools from statsmodels.regression.quantile_regression import QuantReg for i in range(5000): mod = QuantReg(endog.iloc[i].values, tools.add_constant(exog.values)) res = mod.fit(q=0.5).params[1]

Execution time is ~106s ------------=----------- However, using quantile regression (rq function) in R, only spent user system elapsed 6.249 0.374 3.770 --------=----------- Is there any method to reduce the execution time of QuantReg in Python to be comparable to R?