Update tensor_prod.md #464
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #464 +/- ##
==========================================
+ Coverage 78.00% 78.12% +0.12%
==========================================
Files 23 23
Lines 3155 3159 +4
==========================================
+ Hits 2461 2468 +7
+ Misses 694 691 -3 ☔ View full report in Codecov by Sentry. |
|
@sathvikbhagavan and @ChrisRackauckas |
|
For the benchmarks, I was thinking each benchmark problem should have three sections:
Can you do the same for As for the |
|
@sathvikbhagavan. |
|
@ChrisRackauckas and @sathvikbhagavan |
| # Tabular Representation of all Surrogates and their MSE Scores | ||
|
|
||
| | Surrogate Model | MSE Score | | ||
| |-------------------|----------------------| | ||
| | Kriging | 4.70493378010316e-5 | | ||
| | Lobachevsky | 7.967792682690972e-5| | ||
| | Radial Basis | 0.004972603698976124 | | ||
| | RandomForest | 0.2023233139232778 | | ||
| | Poly1 | 0.4124881232761028 | | ||
| | Poly2 | 0.42166909818265136 | | ||
|
|
||
|
|
||
| This structure provides a framework for generating data, training various surrogate models, evaluating their performance on test data, and reporting the best surrogate based on performance metrics like MSE. Adjustments can made to suit the specific evaluation criteria or additional surrogate models. |
There was a problem hiding this comment.
This doesn't make any sense here. What does this have to do with the tensor product function?
There was a problem hiding this comment.
Actually, this is what we can do for all the benchmarks , and this was suggested by @sathvikbhagavan that we should analyse all the models and choose the best one for a specific benchmark.
There was a problem hiding this comment.
benchmarks wouldn't go into the documentation, that would go into the SciMLBenchmarks.
There was a problem hiding this comment.
Also, the solution shouldn't be hardcoded but instead generated by computation
There was a problem hiding this comment.
Generated as in using the Documenter.jl??
There was a problem hiding this comment.
oh ok! But then what should I do for this PR? I mean we can still include choosing the best surrogate for a specific benchmark right?
There was a problem hiding this comment.
Also I am not very much aware of how do we use Weave? You have any tutorial or something??
There was a problem hiding this comment.
just look at the repo instructions https://github.com/SciML/SciMLBenchmarks.jl?tab=readme-ov-file#contributing
There was a problem hiding this comment.
oh ok! And what should I do with this PR? should I generate the code blocks using weave right now?
There was a problem hiding this comment.
Make a weave jmd and open a PR to SciMLBenchmarks
|
@ChrisRackauckas just to clarify, should we have benchmarks section in Surrogates documentation or completely move them to SciMLBenchmarks? |
|
Yes, we should make a new surrogates section |
I have made a new PR for the previous change. This is just changing the tensor_prod function in a more readable form which I think I, @ChrisRackauckas and @sathvikbhagavan agreed. Kindly please review it for merging.
Next I am working on explaining the mathematical concepts for Kriging and other surrogate models. I already gave a sample explaination for Kriging and asked if we could go with the same format.
So in future I will hit a draft PR where I will constantly modify the tutorials section.
Well this is the explaination for Kriging: