Hello all: I’m running into strange behavior with neural ODEs defined with composite functions and FastChain definitions. In short, the below code works w/o issue for small neural networks, but fails for larger networks because of a getindex() issue. See this toy minimal (non)working example: usin…

ForwardDiffSensitivity wasn’t compatible with u0 as a scalar (we always used arrays!). Fixed that here: forwarddiffsensitivity scalar SciML:master ← SciML:forwarddiffsensitivity_scalar opened 01:31…

Neural ODE works for small networks, but throws error for larger networks (getindex() method)?

Specific Domains Machine Learning

ChrisRackauckas August 4, 2022, 2:43pm 4

ForwardDiffSensitivity wasn’t compatible with u0 as a scalar (we always used arrays!). Fixed that here:

Other adjoints will throw a nice error, but instead of fixing the error message here, this dispatch was easy to just solve.

Topic		Replies	Views
Neural ODE in DiffEqFlux that is not a time series Machine Learning diffeq	6	1437	March 30, 2019
DiffEqFlux: neural_ode stops prematurely General Usage diffeq	4	583	January 24, 2019
Parameter tuning in neural ODE Machine Learning diffeq , flux	15	2077	July 21, 2019
A simplified example for DiffEqFlux New to Julia	1	430	July 27, 2021
DiffEqFlux NaN Bug with Neural ODE: "BoundsError: attempt to access 1-element Vector{Float64} at index [2]" Machine Learning neural-network	3	532	October 30, 2021

Neural ODE works for small networks, but throws error for larger networks (getindex() method)?

Related topics