Collection of matlab scripts to run computations on a cluster
Objective-C Matlab
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
README
TODO
matluster_addJobToQueue.m
matluster_collect.m
matluster_findIndex.m
matluster_generateStringFromOptions.m
matluster_insertResult.m
matluster_reshapeResults.m

README

MATLUSTER: MATlab on the CLUSTER

Matluster is a loose collection of scripts that might prove useful for
grid-search and hyper-parameter evaluations on a cluster. It also
features a working example, that shows some of the best practices when working
with Matlab on a cluster.

The scripts work great for simple hyper-paramter evaluation or different
train/test data splits often encountered in machine learning. This toolbox
is utterly useless if you are looking into more complex distributed
Matlab computations, such as map-reduce.

Think of Matluster as follows: You have a working Matlab algorithm that runs
in reasonable time (say 5 hrs) for a specific configuration. You now want to
run the very same algorithm for 1000 different configurations.