Hi, I have a plain GMM package under development at https://github.com/davidavdav/GMMs.git .
I see now that I need to update the Readme.md a bit, I've split off some stuff that was too specific for automatic speaker recognition. The GMMs can now also be full covariance. There is support for EM training with data amounts larger than would fit in (cluster) memory through BigData.jl. The latter package was written because I could not get DArray or memory-mapped i/o to work efficiently on the cluster. You might build a HMM on top of that. ---david On Saturday, October 4, 2014 6:31:10 AM UTC+2, Arshak Navruzyan wrote: > > I'm looking for a functioning HMM implementation. This seems to have a > dependency on MixtureModels which doesn't seem to be maintained. > > https://github.com/nfoti/hmm.jl/blob/master/src/hmm.jl > > Is there a different package (or perhaps this can be used with GMMs > package instead?) > > Thanks. >
