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
Yes, it is a duplicate one. It occurs quite often with iterscheme=:fixed, not for iterscheme=:diffgauge. Does the later (iterscheme=:diffgauge) implement the linsolve for AD?
With the default settings of PEPSKit, the test for tf_ising.jl passes, but failed when I renormalize the tensors before ctmrg svd for projectors. I am confused as I think the normalization improves the stability.
https://github.com/Jutho/KrylovKit.jl/blob/a39cc4f88f5f7a0935aae52a60372b1d0825773b/ext/KrylovKitChainRulesCoreExt/svdsolve.jl#L238C1-L239C24
When running PEPSKit.jl, an bound error occurs here
Does
Ws
here always have length ofn
? If I changefor i in 1:n
tofor i in 1:length(Ws)
, is there any consequence?The text was updated successfully, but these errors were encountered: