You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Both the eval and hyperopt commands can now perform suggest operations in parallel on multiple CPUs. When training the nn_ensemble, running the documents through the source projects is the main bottleneck. We could speed it up by implementing parallel suggest operations there as well. There should probably be a new jobs parameter for the train CLI command specifying the number of parallel jobs, just like for eval and hyperopt.
The text was updated successfully, but these errors were encountered:
Both the
eval
andhyperopt
commands can now performsuggest
operations in parallel on multiple CPUs. When training the nn_ensemble, running the documents through the source projects is the main bottleneck. We could speed it up by implementing parallel suggest operations there as well. There should probably be a newjobs
parameter for thetrain
CLI command specifying the number of parallel jobs, just like foreval
andhyperopt
.The text was updated successfully, but these errors were encountered: