search

Zymrael / awesome-neural-ode

A collection of resources regarding the interplay between differential equations, deep learning, dynamical systems, control and numerical methods.
MIT License
1.3k stars 146 forks source link

Add DiffEqFlux and NeuralNetDiffEq #1

Closed ChrisRackauckas closed 4 years ago

ChrisRackauckas commented 4 years ago

These are packages which solve neural differential equations or use neural differential equation architectures. Supports neural ODEs, neural SDEs, neural DDEs, neuarl PDEs, etc., with GPU support, O(1) backprop, stiff solvers, automatic stiffness detection, etc. (all of the goodies).

Zymrael commented 4 years ago

Sorry for the late response. I am aware of your amazing work on JuliaDiffEq (among other things) and am trying to get familiar with it enough to explore some of my research ideas. Unfortunately, as was also pointed out in the JuliaDiffEq blog post, the limited solver selection of torchdiffeq is likely to hurt / slow down research in this area and I have experienced this first-hand in some of my experiments. Merged & wish you the best of luck in your projects.