How to pass additional arguments to odeint(.) from torchdiffeq library

1

Hi,
I have question about the pytorch library torchdiffeq (https://arxiv.org/pdf/1806.07366.pdf). This is a pytorch implementation of ordinary differential equations where the differential function is defined by a neural network. For example,
dy/dt = f(t,y,\theta), here f(.) is a neural network. My question is if my \theta comes from another neural network and I want to pass it to odeint(.) from the library how should I do that?

Any help is really appreciated.

Thanks

2

Hi,

This is not a library provided by the core developers so I don’t know.
Maybe the people that wrote that lib are reading this forum and will answer. Otherwise, maybe you should open an issue on the repo directly: https://github.com/rtqichen/torchdiffeq

1 Like