Currently the rhs function is defined using sympy and then compiled using numba.
Now that aesara has a numba backend we can use aesara to compile this function instead, which should speed up the compilation of larger models a lot and also makes it easier to define the ODE, because we follow numpy more closely.
Currently the rhs function is defined using sympy and then compiled using numba.
Now that aesara has a numba backend we can use aesara to compile this function instead, which should speed up the compilation of larger models a lot and also makes it easier to define the ODE, because we follow numpy more closely.