I have the same problem like described at Ray[Tune] BayesOpt fails to evaluate more than 11 trials · Issue #28063 · ray-project/ray · GitHub
Is there any solution to automatically stop the optimization?
I have multiple data souces and hat to do a hyper param optimization. One source is a dummy data set and therefor does only need 8-12 tries. All other do need more than 100. It is kind of strange to always check if ray crashed or the server crashed or the optimizer just do not find any better sets.
It would be great to have some kind of "If the optimizer tries 10 x max sample without finding a better set-> then stop the optimizer. "