LogisticRegressionClassifier out of memory

Open Dadibom opened this issue 5 years ago • 2 comments

Trying to train LogisticRegressionClassifier with a large dataset. LogisticRegressionClassifier.train() causes the heap to run out of memory.

The data in question is too large to be able to simply increase the heap size, so I am looking for a solution to train in batches. However, I cannot find any references in the documentation to this.

What is the intended way to train a classifier in batches?

Oct 23 '20 08:10 Dadibom

Please any solution to this yet?

Oct 28 '20 07:10 SirPhemmiey

You could try training in parallel with the functions implemented in classifier_train_parallel.

Mar 01 '24 11:03 Hugo-ter-Doest