Closed StatguyUser closed 6 years ago
This is a classic out of memory error. The code there is already using memory-optimized data structures, so I suggest you reduce the amount of data (use a sample only), reduce the number of projections, number of neighbors, etc. - anything that reduces memory. Also make sure you use all your memory (by default Java will only use 25% for one process), and maybe just get more memory if you have that much data.
Closing: This is not a bug in ELKI.
how do i assign more memory for JAVA? is there any option in ELKI?
The memory limit must be set for the Java JVM, it cannot be set by the application.
The Java option is usually called -Xmx
, but YMMV.
This applies to any Java application. E.g., https://docs.oracle.com/javase/8/docs/technotes/tools/windows/java.html
I am trying to cluster word2vec vectors which came from text documents. These are 15 decimal numbers. I tried using DBSCAN, fastoptics etc, however i get below error. Can anyone help me on this? I tried using parser-vector-type as
SparseFloatVector
,FloatVector
and the default one too, however i end up getting below error every time