# Optimization v3.16.0 errors in precompile

**URL:** <https://discourse.julialang.org/t/optimization-v3-16-0-errors-in-precompile/103229>\
**Category:** Optimization (Mathematical)\
**Tags:** package\
**Created:** [August 26, 2023, 6:35pm UTC](https://discourse.julialang.org/t/optimization-v3-16-0-errors-in-precompile/103229 "2023-08-26T18:35:59Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![jaksaa](https://avatars.discourse-cdn.com/v4/letter/j/91b2a8/32.png) [@jaksaa](https://discourse.julialang.org/u/jaksaa)\
**Post date:** [August 26, 2023, 6:35pm UTC](https://discourse.julialang.org/t/optimization-v3-16-0-errors-in-precompile/103229/1 "2023-08-26T18:35:59Z")

</div>

Have  
Julia Version 1.9.2  
Optimization v3.16.0

Two errors are present  
`│ ┌ Error: Error during loading of extension OptimizationSparseFinitediffExt of Optimization, use `Base.retry_load_extensions()` to retry.`

and  
`│ ┌ Error: Error during loading of extension OptimizationSparseForwarddiffExt of Optimization, use `Base.retry_load_extensions()` to retry.`

Then did  
`test Optimization`

This resulted in many warnings of deprecated functions e.g.

`┌ Warning: `ADAM` is deprecated, use `Adam` instead. │ caller = top-level scope at ADtests.jl:449 └ @ Core ~/.julia/packages/Optimization/HPEh3/test/ADtests.jl:449`

As well as errors, the two above and this one:

`DiffEqFlux: Test Failed at ~/.julia/packages/Optimization/HPEh3/test/diffeqfluxtests.jl:113`  
Expression: result\_neuralode2.objective \< 10  
Evaluated: 10.487717f0 \< 10

Then it stops testing.  
Is this error expected?  
Is it related to the precompile errors?

---

<div class="post-metadata">

**Author:** ![ChrisRackauckas](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/chrisrackauckas/32/77_2.png) [@ChrisRackauckas](https://discourse.julialang.org/u/ChrisRackauckas)\
**Post date:** [August 27, 2023, 1:04pm UTC](https://discourse.julialang.org/t/optimization-v3-16-0-errors-in-precompile/103229/2 "2023-08-27T13:04:34Z")

</div>

> [@jaksaa](#):
>
> `DiffEqFlux: Test Failed at ~/.julia/packages/Optimization/HPEh3/test/diffeqfluxtests.jl:113`  
> Expression: result\_neuralode2.objective \< 10  
> Evaluated: 10.487717f0 \< 10
> 
> Then it stops testing.  
> Is this error expected?  
> Is it related to the precompile errors?

That’s unrelated. That test can randomly error sometimes due to bad initializations of the neural network. I wouldn’t be too worried aobut that.

---

<div class="post-metadata">

**Author:** ![jaksaa](https://avatars.discourse-cdn.com/v4/letter/j/91b2a8/32.png) [@jaksaa](https://discourse.julialang.org/u/jaksaa)\
**Post date:** [August 31, 2023, 7:36pm UTC](https://discourse.julialang.org/t/optimization-v3-16-0-errors-in-precompile/103229/3 "2023-08-31T19:36:24Z")

</div>

Thank you for your response.  
Does this mean the error threshold is too tight?  
Is there something I can do to improve the initialization?

---

<div class="post-metadata">

**Author:** ![ChrisRackauckas](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/chrisrackauckas/32/77_2.png) [@ChrisRackauckas](https://discourse.julialang.org/u/ChrisRackauckas)\
**Post date:** [August 31, 2023, 8:35pm UTC](https://discourse.julialang.org/t/optimization-v3-16-0-errors-in-precompile/103229/4 "2023-08-31T20:35:10Z")

</div>

> [@jaksaa](#):
>
> Does this mean the error threshold is too tight?

Probably. We should probably make the test use multiple initializations or something to make it more robust.