Open andrewfullard opened 3 weeks ago
Your tests look reasonable, I think checking for errors like NaN, inf etc is a good use of them for this kind of machine learning work. I think the workflow itself, when you connect the parts together, can get its own tests that will probably be a little easier to write.
Could you please take a look at the tests I have developed up to now? I have considered a separate branch for the tests. Moreover, the test jobs run in parallel.
I have a few questions about building a connection between the outputs of the optimization, which are going to be used as the input of the sensitivity analysis. Hopefully, we can discuss it before tomorrow's class so I can have better vision to make more progress toward developing a reasonable workflow.
Your tests look reasonable, I think checking for errors like NaN, inf etc is a good use of them for this kind of machine learning work. I think the workflow itself, when you connect the parts together, can get its own tests that will probably be a little easier to write.