DataDrivenDiffEq.jl icon indicating copy to clipboard operation
DataDrivenDiffEq.jl copied to clipboard
verified publisher icon SciML

Reliability of OccamNet Tests

Open AlCap23 opened this issue 4 years ago • 1 comments

It seems there are some issues with the tests of OccamNet since recently, probably related to infeasible combinations of functions ( e.g. 1 / 0 , log(-10.0) or similar calls ) . This should be a reasonable quick fix, but I open this issue as a reminder and because I will probably have time for next week.

AlCap23 avatar Jan 16 '22 16:01 AlCap23

Just some tasks related

  • [ ] Add more tests to src\symbolic_regression/occamnet.jl, see here
  • [ ] I am running 100 training loops locally right now. There are no errors using v1.7.1 on the current MacOs. Why do the tests fail on the CI

AlCap23 avatar Jan 17 '22 09:01 AlCap23