I am trying to run train HMM using train function with multiple cores but it would use only single code.
bw <- train(hmm, seq, method = "BaumWelch", maxiter=1000, logspace = FALSE, cores="autodetect", quiet=TRUE)
As I was reading the code from the package cores="autodetect" should initialise n-1 cores to work on.
Could you help me to understand this and how to enable using multiple cores.
I would also would like to know number of iterations it took to converge. I know quiet=FALSE will give me feedback but user would have to manually look for it, I managed to get if the training converges or not by catching the warning message to boolean variable but couldn't get the number of iterations store to a variable.
Hi @shaunpwilkinson,
I am trying to run train HMM using train function with multiple cores but it would use only single code. bw <- train(hmm, seq, method = "BaumWelch", maxiter=1000, logspace = FALSE, cores="autodetect", quiet=TRUE) As I was reading the code from the package cores="autodetect" should initialise n-1 cores to work on. Could you help me to understand this and how to enable using multiple cores.
I would also would like to know number of iterations it took to converge. I know quiet=FALSE will give me feedback but user would have to manually look for it, I managed to get if the training converges or not by catching the warning message to boolean variable but couldn't get the number of iterations store to a variable.
Cheers, Rahil