solverstoppingjulia / stopping.jl Goto Github PK
View Code? Open in Web Editor NEWA framework to implement iterative algorithms
License: Other
A framework to implement iterative algorithms
License: Other
type SlackModel has no field counters
Stacktrace:
[1] getproperty(x::NLPModelsModifiers.SlackModel{Float64, Vector{Float64}, MathOptNLPModel}, f::Symbol)
@ Base ./Base.jl:33
[2] _resources_check!
@ ~/.julia/packages/Stopping/I991u/src/Stopping/NLPStoppingmod.jl:311 [inlined]
[3] _main_pb_check!
@ ~/.julia/packages/Stopping/I991u/src/Stopping/GenericStoppingmod.jl:541 [inlined]
[4] stop!(stp::NLPStopping{FletcherPenaltyNLP{Float64, Vector{Float64}, Val{2}, Float64, FletcherPenaltyNLPSolver.IterativeSolver{Float64, Vector{Float64}, Krylov.LsqrSolver{Float64, Vector{Float64}}, Krylov.LnlqSolver{Float64, Vector{Float64}}, Krylov.MinresSolver{Float64, Vector{Float64}}}}, StoppingMeta{Float64, Float64, Nothing, Int64}, StopRemoteControl, NLPAtX{Float64, Vector{Float64}, Matrix{Float64}}, NLPStopping{NLPModelsModifiers.SlackModel{Float64, Vector{Float64}, MathOptNLPModel}, StoppingMeta{Float64, Vector{Float64}, Nothing, Int64}, StopRemoteControl, NLPAtX{Vector{Float64}, Vector{Float64}, Matrix{Float64}}, VoidStopping{Any, StoppingMeta, StopRemoteControl, GenericState, Nothing, VoidListofStates}, VoidListofStates}, VoidListofStates}; no_opt_check::Bool, kwargs::Base.Iterators.Pairs{Union{}, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
@ Stopping ~/.julia/packages/Stopping/I991u/src/Stopping/GenericStoppingmod.jl:374
[5] stop!
@ ~/.julia/packages/Stopping/I991u/src/Stopping/GenericStoppingmod.jl:347 [inlined]
when initializing
stp = NLPStopping(nlp, StoppingMeta(), StopRemoteControl(), GenericState(nlp.meta.x0))
Is this the correct behavior or we should return something else?
It would be great to set up some tools to uniformize coding here.
Currently it is only available in the complete constructor (with pb, meta, state)
I am suspecting, we will discover a couple of typing issues (maybe in the StoppingMeta
for starter)
We started in #64 for the admissible functions
In nlp_admissible_functions.jl, check that the current_score
does not convert to Float64
because of the 0.0
.
Note that currently we use max_cntrs
https://github.com/vepiteski/Stopping.jl/blob/d990b6ef4541b20446275d3c26b17ba83ef00ac2/src/Stopping/StoppingMetamod.jl#L90
Aim: easily do what others do.
nlp_admissible_functions.jl#L12
Maybe add a keyword with an unbounded threshold.
in the case where only one status is returned.
This issue is used to trigger TagBot; feel free to unsubscribe.
If you haven't already, you should update your TagBot.yml
to include issue comment triggers.
Please see this post on Discourse for instructions and more details.
If you'd like for me to do this for you, comment TagBot fix
on this issue.
I'll open a PR within a few hours, please be patient!
nlp_admissible_functions.jl#L12
The signature of the function norm
is Real
not Float64
.
It is an NLPStopping
that behaves like an AbstractNLPModel
, it would
x
?This would help using non-stopping solvers.
Same if objective function is Inf
and we minimize. @vepiteski
using ADNLPModels, Stopping
nlp = ADNLPModel(x -> 100 * (x[2] - x[1]^2)^2 + (x[1] - 1)^2, [-1.2; 1.0])
state = NLPAtX(nlp.meta.x0)
state.x .= 1 # also modifies nlp.meta.x0
Related to JuliaSmoothOptimizers/AdaptiveRegularization.jl#87
Add options that woul be useful for the users in the meta.
#92 we can set minimize = false only in the next release of ADNLPModels.
These getter return nothing
for AbstractStopping
, in case someone redefines a Stopping
.
need to add a stalled_tol
in Meta.
So there are "three" different states in the list.
Calling stop!
when the current_score
contains NaN
but the objective function is -Inf
, which should be a good news :).
@vepiteski
The last update of ADNLPModels
doesn't seem to be compatible with our unit tests.
Two options:
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.