You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Would it be possible to add padding support to convolutions? My current application relies on keeping the input size fixed but uses a very small model (see below). (Without padding SimpleChains outperforms Lux by >30x on these workloads)
model =Chain(
Conv((11,), 1=>16, Lux.relu; pad=Lux.SamePad()),
Conv((11,), 16=>1; pad=Lux.SamePad()),
)
Would it be possible to add padding support to convolutions? My current application relies on keeping the input size fixed but uses a very small model (see below). (Without padding SimpleChains outperforms Lux by >30x on these workloads)
cc @ChrisRackauckas (this is for the learned solvers work)
The text was updated successfully, but these errors were encountered: