search

google-research / torchsde

Differentiable SDE solvers with GPU support and efficient sensitivity analysis.
Apache License 2.0
1.56k stars 195 forks source link

sdeint_adjoint for f_and_g_prod type SDE, how to set parameter? #146

Open chunqishi opened 4 months ago

chunqishi commented 4 months ago

I got the following errors:

File ".venv\lib\site-packages\torchsde_core\adjoint.py", line 271, in sdeint_adjoint extra_solver_state = solver.init_extra_solver_state(ts[0], y0) File ".venv\lib\site-packages\torchsde_core\methods\reversible_heun.py", line 59, in init_extra_solver_state return self.sde.f_and_g(t0, y0) + (y0,) File ".venv\lib\site-packages\torchsde_core\base_sde.py", line 92, in f_and_g_default return self.f(t, y), self.g(t, y) File ".venv\lib\site-packages\torchsde_core\base_sde.py", line 85, in g_default raise RuntimeError("Method g has not been provided, but is required for this method.") RuntimeError: Method g has not been provided, but is required for this method.