java OutOfMemoryError: heap space

[yuliagrishina]

Hi Greg, I keep getting java OutOfMemoryError when I try to train a new model. I already switched to COREF_TRAIN_PREDICT mode as you suggest in the readme file, but the error is still there. Increasing the heap size does not help. Any suggestions how to fix that would be very much appreciated! Thanks! Yulia

UPD: I found the solution. it happened because my conll file was somehow missing document borders and was read in as a single document. adding document borders ("#begin document..") solved the problem.

[gregdurrett]

Hi Yulia,

Can you send me the output log? The trick I suggested only works when it's crashing due to the coreference scorer. When you're training, you still need to load the entire train set and keep it in memory. How much data are you training on and what size heap are you providing?

Greg

On Tue, Apr 26, 2016 at 2:03 AM, yuliagrishina [email protected] wrote:

Hi Greg, I keep getting java OutOfMemoryError when I try to train a new model. I already switched to COREF_TRAIN_PREDICT mode as you suggest in the readme file, but the error is still there. Increasing the heap size does not help. Any suggestions how to fix that would be very much appreciated! Thanks! Yulia

— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub https://github.com/gregdurrett/berkeley-entity/issues/10