# How to benchmark my loss function with sciml\_train using different optimizers?

**URL:** <https://discourse.julialang.org/t/how-to-benchmark-my-loss-function-with-sciml-train-using-different-optimizers/47852>\
**Category:** General Usage\
**Tags:** question, diffeq, sciml\
**Created:** [October 6, 2020, 12:49pm UTC](https://discourse.julialang.org/t/how-to-benchmark-my-loss-function-with-sciml-train-using-different-optimizers/47852 "2020-10-06T12:49:12Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Alec\_4](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/alec_4/32/18466_2.png) [@Alec\_4](https://discourse.julialang.org/u/Alec_4)\
**Post date:** [October 6, 2020, 12:49pm UTC](https://discourse.julialang.org/t/how-to-benchmark-my-loss-function-with-sciml-train-using-different-optimizers/47852/1 "2020-10-06T12:49:12Z")

</div>

Does anyone know some good way to benchmark my loss function trained with sciml\_train using different optimizers? Just like the way that WorkPrecisionSet() can do to show the performance differences due to using different algorithms as below:

```julia
setups = [Dict(:alg=>BS3()),
          Dict(:alg=>Tsit5()),
          Dict(:alg=>RK4()),
          Dict(:alg=>DP5()),]
wp = WorkPrecisionSet(prob,abstols,reltols,setups;
                      appxsol=test_sol,maxiters=Int(1e5),error_estimate=:L2)
plot(wp)

```

Is there some suitable function to treat my loss as argument to do so?  
By the way, if you are familiar with sciml\_train, could you also recommend me some other fast optimizers like ADAM?

---

<div class="post-metadata">

**Author:** ![Alec\_4](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/alec_4/32/18466_2.png) [@Alec\_4](https://discourse.julialang.org/u/Alec_4)\
**Post date:** [October 8, 2020, 8:04am UTC](https://discourse.julialang.org/t/how-to-benchmark-my-loss-function-with-sciml-train-using-different-optimizers/47852/2 "2020-10-08T08:04:30Z")

</div>

😕