SciML / DiffEqFlux.jl

Pre-built implicit layer architectures with O(1) backprop, GPUs, and stiff+non-stiff DE solvers, demonstrating scientific machine learning (SciML) and physics-informed machine learning methods
https://docs.sciml.ai/DiffEqFlux/stable
MIT License
871 stars 157 forks source link

NeuralODE problem with component arrays not working #497

Closed yewalenikhil65 closed 3 years ago

yewalenikhil65 commented 3 years ago

Hi , using following packages

  [b0b7db55] ComponentArrays v0.8.19
  [aae7a2af] DiffEqFlux v1.34.0
  [587475ba] Flux v0.11.6
  [429524aa] Optim v1.2.4

the example of NeuralODE using componentarrays mentioned here gives following error..

image

ChrisRackauckas commented 3 years ago

This should be an issue with ComponentArrays?

yewalenikhil65 commented 3 years ago

This should be an issue with ComponentArrays?

yes i guess.. i raised the issue here since I first found it at https://diffeqflux.sciml.ai/stable/Flux/#Using-ComponentArrays-for-neural-network-layers-1 As i was unsure what is real issue, i raised issue at both repos.. diffeqflux and componentarrays

yewalenikhil65 commented 3 years ago

fixed by ComponentArrays maintainers with their new update yesterday