pred <- predict(Theta1, Theta2, X[rp[1],]) Error in rep(0, m) : invalid 'times' argument

[rajasekhar-thiru]

Hi Fared, The 94th line of the ex3_nn.R reads as follows:
pred <- predict(Theta1, Theta2, X[rp[i],])

It is called from inside this loop: for (i in 1:m){ pred <- predict(Theta1, Theta2, X[rp[i],]) }

I am getting the following error:

pred <- predict(Theta1, Theta2, X[rp[1],]) Error in rep(0, m) : invalid 'times' argument

The predict function worked fine when it was invoked the first time in the same program at line 76 pred <- predict(Theta1, Theta2, X)

I could not find anything wrong with the repeat command used inside the predict function at row 14 p <- rep(0,m) where m <- dim(X)[1]

I am not sure if the function results are cached in R the first time which causes problem when called again.

Can you please help.

https://github.com/faridcher/machine-learning-course/blob/master/Solutions/mlclass-ex3/ex3_nn.R

[faridcher]

The problem could be from the X. what is the value of X? Are you trying my solutions or yours? Debug the code and check for the value of X and then m <- dim(X)[1].

hope it helps.

[rajasekhar-thiru]

Farid, I am trying your solution... X is a 5000 by 400 matrix.

The predict function is called twice in the program. It worked fine first time and aborts with this error 2nd time.

Thanks

Raja

---

On Sat, 5/13/17, Farid Cheraghi notifications@github.com wrote:

Subject: Re: [faridcher/machine-learning-course] pred <- predict(Theta1, Theta2, X[rp[1],]) Error in rep(0, m) : invalid 'times' argument (#3) To: "faridcher/machine-learning-course" machine-learning-course@noreply.github.com Cc: "rajasekhar-thiru" rajasek@yahoo.com, "Author" author@noreply.github.com Date: Saturday, May 13, 2017, 3:51 PM

The problem could be from the X. what is the value of X? Are you trying my solutions or yours?

Debug the code and check for the value of X and then m <- dim(X)[1]. hope it helps.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[faridcher]

I just opened the mlclass-ex3.Rproj and sourced the ex3_nn.R and it runs without errors.

predict function could be called many times and R doesn't cache the function calls.

[rajasekhar-thiru]

thanks. I will continue looking line by line...