Charles Smith

I find that the code also takes a really long time to run using a multvariate normal distribution ``` #User Input N = 7 tol = 2 alg = HCubatureJL()...

I'm finding that with CubaSUAVE and CubaVegas that for large N (N > 10) the solutions are inaccurate. Is this expected behavior or am I doing something wrong? ``` #User...

Alternatively, I can try to parallelize the SDE paths by using a pmap or map loop, but that seems to run into similar issues: Example code: ``` ### Using Packages...

I copied the example you gave and it works just fine, but when I apply it to my problem I still seem to get errors: MethodError: no method matching (::var"#46#47")(::Array{Float64,1},...

Awesome, thank you so much! One last question: How do I get the schiml_train function to update the intial position of the SDE (u0 in the example). My first guess...

Yes that works for me. I’ll look through all your comments on the code this evening and implement them. I’ll also send results from running the code this evening as...

Thanks for the quick response. That is exactly what I want. However, when I make that change that you suggest, the sciml_train only updates the u0 parameters not the p...

I made your adjustment and now sciml_train updates neither the parameters of the NN nor the initial positions. Any ideas? [Normal_SDEs_NN_v3.txt](https://github.com/SciML/DiffEqFlux.jl/files/4381869/Normal_SDEs_NN_v3.txt) Thank you for all the help so far.

Okay. Thank you for all the help. I made both of those adjustments, and still sciml_train updates neither the parameters of the NN nor the initial positions. Any ideas? ```...

Hello, I'm just checking in on this to see if you have thought about it and if you have ideas to solve this issue. Thank you!