rl-institut / smooth

Simulation framework for non-linear state-dependant energy systems
Other
6 stars 7 forks source link

no cache in run_optimization #169

Closed LauWien closed 4 years ago

LauWien commented 4 years ago

After the optimization algorithm is finished it would be awesome if the resulats are stored even though post_processing is not finished. Also storing each new gradient descending result. So if an error occures or you need to stop the process you will still have some results