-
MWE:
```julia
using DiffEqFlux, Flux, Optim, OrdinaryDiffEq, Plots
using LinearAlgebra
function rober(du,u,p,t)
y₁,y₂,y₃ = u
k₁,k₂,k₃ = p
du[1] = -k₁*y₁+k₃*y₂*y₃
du[2] = k₁*y₁-k₂…
-
**Problem:** In Stochastic Gradient Descent sometimes the descent will suddenly go unstable even with a step size set as low as you can go (limited by FLoat32 precision). This is not preventable wit…
cems2 updated
4 years ago
-
Hi!
I'm trying to run the Nerual ODE demo code and have run into this issue.
Specifically, when I run the demo code under the section `Training a Neural Ordinary Differential Equation` (exact cod…
-
Looking through [`neural_de.jl`](https://github.com/JuliaDiffEq/DiffEqFlux.jl/blob/master/src/neural_de.jl) it doesn't look like it'll be hard to implement a `NeuralDAE`.
I'm interested in using th…
-
This issue is superficially similar to #81 but the resolution there was to use a newer Julia. However, this happens in the latest available stable Julia (v1.3).
**Problem**
In julia 1.3 on MacOS…
cems2 updated
4 years ago
-
In a NeuralODE problem I am testing two trivial models
```
model1 = FastDense(2,2)
model2 = FastChain(model1)
```
I believe these are nominally identical models. Yet when called in a batch mode,…
cems2 updated
4 years ago
-
Here is an 1D problem where the model is trying to learn the function `f(u)=u.^3`. The data is a `(1,200)`-dimensional array, where 200 is the batch size.
```
using Flux, DiffEqFlux, OrdinaryDiffE…
-
First off, awesome paper, thank you!
When I try to run the NeuralODE example on CPU, it works great. However, when I try to switch it over to the GPU using the approach described in the paper i.e.,…