Sathvik Bhagavan

Hello, Can you provide a code snippet where it is failing?

Scaling the timeseries does work but it can easily break: ```julia tsteps = 0:0.1:1 data = reshape( sin.(15*tsteps),1,:) scale = 10 du,u = collocate_data(data,scale*tsteps,EpanechnikovKernel()) ``` Error: ```julia ERROR: LinearAlgebra.SingularException(2) Stacktrace:...

@ChrisRackauckas can this issue be closed as https://github.com/SciML/DiffEqFlux.jl/pull/853 addresses bandwidth being passed by the user or should we also implement point 3 mentioned above: "Implement signal filters such that the...

Is this still an issue?

Actually I thought as `PhysicsInformedNN` works with GPUs, this wouldn't be a problem, but apparently not 😅 MWE: This is using a branch https://github.com/SciML/NeuralPDE.jl/tree/sb/try_gpu_nnode (had to do some fixes) ```julia...

I think this can be closed as well with #802 merged.

There are couple of PRs related to it - https://github.com/SciML/NeuralPDE.jl/pull/673 & https://github.com/SciML/NeuralPDE.jl/pull/506. Not sure if it is still relevant.

Hi @MilesCranmer, I wanted to look into it if thats alright and had a few questions: How does the algorithm work with parametric equations? Iiuc, we can pass in some...

@ChrisRackauckas can you review this?

>What is this doing that cumsum after the solver does not? So, the aim is to get back an integrated time series back from a time series. Basically collecting intermediate...