search

JuliaTrustworthyAI / LaplaceRedux.jl

Effortless Bayesian Deep Learning through Laplace Approximation for Flux.jl neural networks.
https://www.taija.org/LaplaceRedux.jl/
MIT License
39 stars 3 forks source link

check if the chain ends with a softmax/sigmoid layer and accordingly modify the predict functions. #98

Closed pasq-cat closed 2 months ago

pasq-cat commented 3 months ago

as the title say, the predict function should know when the chain already apply a softmax/sigmoid layer to avoid applying it 2 times in a row.

pat-alt commented 2 months ago

I've edited the title since the main thing is being fixed by #99

pasq-cat commented 2 months ago

@pat-alt should we close this issue?

pat-alt commented 2 months ago

Yup!