Better BPINN ode Solver #853
Conversation
|
@ChrisRackauckas @Vaibhavdixit02 for the new formulations, merging of this PR first and PR #842 later is recommended. |
test/bpinnexperimental.jl
Outdated
| @@ -0,0 +1,140 @@ | |||
| using Test, MCMCChains | |||
There was a problem hiding this comment.
this new test file is not actually ran.
| @test abs(param1 - p) < abs(0.3 * p) | ||
|
|
||
| #-------------------------- solve() call | ||
| # (lux chain) | ||
| @test mean(abs.(physsol2 .- sol3lux_pestim.ensemblesol[1])) < 0.15 | ||
| @test mean(abs.(physsol2 .- pmean(sol3lux_pestim.ensemblesol[1]))) < 0.15 |
There was a problem hiding this comment.
nope, the mean is required as the solution's standard deviation are different at domain points, sometimes these uncertainties can be large enough for the tests to fail. so i just take the means for testing.
src/collocated_estim.jl
Outdated
| @@ -0,0 +1,46 @@ | |||
| # suggested extra loss function for ODE solver case | |||
| function L2loss2(Tar::LogTargetDensity, θ) | |||
There was a problem hiding this comment.
why is this a separate file/
|
My understanding is that the proposed changes from the last review were not implemented and this was just rebased? Is there a plan to finish this? |
|
This is ready @ChrisRackauckas the other test failures seem unrelated. Let me know if changes are needed. |
Checklist
contributor guidelines, in particular the SciML Style Guide and
COLPRAC.
Additional context
Add any other context about the problem here.