This should show up in my repo only #1
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Fixes SciML/DifferentialEquations.jl#610 Is non-breaking
Setup solve for adjoints to deprecate concrete_solve
More explanatory error message when missing algorithm
safer ReverseDiff
Update License [ci skip]
Use vecvecapply in dense calculate_solution_errors
This changes the norms on dual numbers to include the dual portions. This means that the solution to the ODE under automatic differentiation is not necessarily the same as without. However, SciML/SciMLSensitivity.jl#273 nicely highlights some edge cases that show that this is somewhat necessary. In that case, the integrator has zero error, and so time steps are unaffected by tolerances which causes gradients to explode since they are not tolerance controlled. This change adds them to the summation and the length, making the solution by dual numbers equivalent to the result of forward sensitivity analysis under the default norm. @YingboMa what do you think? @pkofod you might want to test this with Pumas. It should give it some more robustness.
Use JuliaGPU's GitLab CI setup
Co-authored-by: Yingbo Ma <[email protected]>
Co-authored-by: Yingbo Ma <[email protected]>
Co-authored-by: Yingbo Ma <[email protected]>
Check pointer not values when deciding whether to remake
You know how sometimes you later think that some idea was a really really bad idea? Yes, this was one of them. "It's a pair of ODEs, so it's a pair of mass matrices!". Yeah, but that breaks all implicit solver fallbacks, and even when implicit methods specialized to second order ODEs exist, mass_matrix=I will still work as the special case to catch, so it's just better... Anyways, it didn't matter until last week when ArrayPartition suddenly became compatible with all implicit solvers, except for the fact that it would fail because the mass matrix was by default a tuple. So, let's fix this and yay, 2nd order ODEs are now better.
Relax norm test (floating point numbers are really weird)
mass matrix I for DynamicalODEFunction
allow CUDA.jl
Make DiffEqArrayOperator ExponentialUtilities.jl compatible
…6-17-06-10-588-1862310616 CompatHelper: bump compat for "ChainRulesCore" to "0.9"
auto-convert integer inputs
IncrementingODEProblem and IncrementingODEFunction
Remake for DEAlgorithm
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.