https://issues.apache.org/jira/browse/MAHOUT-918 is a project to do this. It is not checked in.
What is the nature of your data? It may be that you can train from a sampled subset of your full corpus. On Tue, Dec 20, 2011 at 11:06 PM, selva <[email protected]> wrote: > Hi everyone, > I modified mahout sgd code for my input format and generated > model for my input. But i have large amount of data, for that i need > map/reduce for sgd algorithm. Kindly i request to give me a solution for > this... > > When will map/reduce release for mahout SGD Classification? > When will mahout 0.6 release ? > > Thanks, > Selva -- Lance Norskog [email protected]
