Sathvik Bhagavan

Yeah, doc build is timing out. Currently checking examples which are slow and trying to speed it up.

The loss functions generated are incorrect. Running this example: ```julia @parameters θ @variables u(..) Dθ = Differential(θ) # 1D ODE eq = Dθ(u(θ)) ~ θ^3 + 2 * θ +...

Superseded by #877, Closing this.

>I'm not sure the regular queue would work because some want to use GPUs? Yes, I was testing out the build times in the normal queue. I have reverted back...

I look at the timestamps of the debug logs. I have fixed one of the tutorial which got stuck for more than 12hrs. Hopefully the build should go through this...

>Which variable corresponds to the phase angle? @HuynhTran0301, can you answer this?

Do something like: (You have to do a forward pass) ```julia ts = 0.0:0.1:25.0 |> collect x = phi[1](ts', res.u.depvar.delta1) plot(ts, x') ``` where `phi` is the vector of `NeuralPDE.phi`,...

Wait, are you using `NNDAE`? I think for `NNDAE`, only `GridTraining` is implemented - https://github.com/SciML/NeuralPDE.jl/blob/master/src/dae_solve.jl#L75 (NNDAE has a diffeq interface, not symbolic). I think other strategies need to be implemented...

Yes, the strategies need to be refactored such that it can be used for both `NNODE` and `NNDAE`. See the strategies present for `NNODE`. Using `PhysicsInformedNN` may not be the...

`NNODE` & `NNDAE` only work with diffeq interface, not symbolic. You would have to write out the equations in a function similar to how you would write for solving it...