[mlpack-git] [mlpack/mlpack] parallel sgd (#603)

ranjan notifications at github.com
Thu Apr 21 15:02:50 EDT 2016


> @@ -39,28 +39,38 @@ double ParallelSGD<DecomposableFunctionType>::Optimize(arma::mat& iterate)
>    double lastObjective = DBL_MAX;
>  
>    //get maximum number of threads that will be running. with is defined by OMP_NUM_THREADS
> -  int num_thread=omp_get_max_threads();
> +  size_t num_thread=omp_get_max_threads();
> +
> +  //T: after T iteration it will go for tollerence checking;  we can define T in many way
> +  //size_t T=maxIterations/num_thread;
> +  //Here I have thaken T=contant
> +  size_t T=1000;

T can be a function of number of Thread , maxIterations, current  Iteration value . Here I have taken Constant.  If T=1. It will check tolerance each time which will slow down the algorithm.

---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/603/files/39c317571e2014f017d648ae72f586c8b0148069..2378633258a12fc751a0265ec13df2824bacb121#r60637451
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160421/af53e0fa/attachment.html>


More information about the mlpack-git mailing list