-
Hi, I tried to train a Neural ODE with Discretecallback with `sensealg=InterpolatingAdjoint(autojacvec=ZygoteVJP(), checkpointing=true)` on GPU, but said:
```julia
Only `ReverseDiffVJP` and `Enzym…
-
-
https://arxiv.org/abs/2002.08071
-
Many graph and node-level problems are fundamentally about diffusion. Information **diffuses** from one node to another. In this way, solutions to these problems can be modeled after heat kernel, whic…
-
# Title
NODEA - Neural Ordinary Differential Equations Anonymous
# Description
This sessions is supposed to be an introduction to neural ordinary differentiable equations (Neural ODEs). As on…
-
I recently came across this paper from last years NeurIPS
https://ghliu.github.io/assets/pdf/neurips-snopt-slides.pdf
https://github.com/ghliu/snopt
I was wondering if there are any plans to …
-
Interesting architecture for consideration for inclusion in this package:
https://arxiv.org/pdf/2105.14969.pdf
-
Implement the ideas from the [Neural ODEs paper](https://arxiv.org/pdf/1806.07366.pdf) as a decoder for the dynamics module.
-
https://arxiv.org/abs/2007.04504 from @jessebett . Have you thought about using TaylorSeries.jl for a Julia implementation?
-
In this [notebook ](https://github.com/DifferentiableUniverseInitiative/flowpm/blob/nbody_ode/notebooks/Simulation-WrongCosmo.ipynb) I try to find the correction terms to compensate for the PM approxi…