could not find function "get_cost"

[tdhock]

Hey @IshmaelBelghazi nice to see that you wrote some demos. However I am getting an error when I try to run them. Maybe you need to export get_cost?

thocking@silene:~/R$ R --vanilla

R version 3.2.1 (2015-06-18) -- "World-Famous Astronaut"
Copyright (C) 2015 The R Foundation for Statistical Computing
Platform: x86_64-unknown-linux-gnu (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

  Natural language support but running in an English locale

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

> library(bigoptim)
Loading required package: Matrix
demo> demo("example_SAG", package="bigoptim")


    demo(example_SAG)
    ---- ~~~~~~~~~~~

Type  <Return>   to start : 

> library(Matrix)

> data(rcv1_train)

> X <- rcv1_train$X

> ##X <- cBind(rep(1, NROW(X), X), X)
> y <- rcv1_train$y

> n <- NROW(X)

> p <- NCOL(X)

> ## Setting seed
> ##set.seed(0)
> maxIter <- n * 20 

> lambda <- 1/n

> tol <- 0

> print("Running Stochastic average gradient with constant step size\n")
[1] "Running Stochastic average gradient with constant step size\n"

> ## -----------------------------------------------------------------------------
> ## SAG with Constant step size
> sag_constant_fit <- sag_fit(X=X, y=y, lambda=lambda, maxiter=maxIter,
+                             tol=0, fit_alg="constant", model="binomial")

> cost_constant <- get_cost(sag_constant_fit, X=X, y=y) 
Error in eval(expr, envir, enclos) : could not find function "get_cost"
> demo("example_SAG2", package="bigoptim")


    demo(example_SAG2)
    ---- ~~~~~~~~~~~~

Type  <Return>   to start : 

> ## Loading Data set
> data(covtype.libsvm)

> ## Normalizing Columns and adding intercept
> X <- cbind(rep(1, NROW(covtype.libsvm$X)), scale(covtype.libsvm$X))

> y <- covtype.libsvm$y

> y[y == 2] <- -1

> n <- NROW(X)

> p <- NCOL(X)

> ## Setting seed
> #set.seed(0)
> ## Setting up problem
> maxiter <- n * 20  ## 10 passes throught the dataset

> lambda <- 1/n 

> tol <- 1e-4

> ## -----------------------------------------------------------------------------
> ## SAG with Constant step size
> print("Running Stochastic Average Gradient with constant step size")
[1] "Running Stochastic Average Gradient with constant step size"

> sag_constant_fit <- sag_fit(X=X, y=y, lambda=lambda, maxiter=maxiter,
+                             tol=tol, family="binomial",
+                             fit_alg="constant", standardize=FALSE)

> cost_constant <- get_cost(sag_constant_fit, X, y)
Error in eval(expr, envir, enclos) : could not find function "get_cost"
In addition: Warning message:
In sag_fit(X = X, y = y, lambda = lambda, maxiter = maxiter, tol = tol,  :
  Optimisation stopped before convergence. Try incrasing maximum number of iterations
> demo("monitoring_training", package="bigoptim")


    demo(monitoring_training)
    ---- ~~~~~~~~~~~~~~~~~~~

Type  <Return>   to start : 

> suppressPackageStartupMessages(library(ggplot2))

> suppressPackageStartupMessages(library(glmnet))

> ## Loading Data set
> data(covtype.libsvm)

> ## Normalizing Columns and adding intercept
> X <- cbind(rep(1, NROW(covtype.libsvm$X)), scale(covtype.libsvm$X))

> y <- covtype.libsvm$y

> y[y == 2] <- -1

> n <- NROW(X)

> p <- NCOL(X)

> ## Setting seed
> #set.seed(0)
> ## Setting up problem
> n_passes <- 50  ## number of passses trough the dataset

> maxiter <- n * n_passes

> lambda <- 1/n 

> tol <- 0

> family <- "binomial"

> fit_algs <- list(constant="constant",
+                  linesearch="linesearch",
+                  adaptive="adaptive")

> sag_fits <- lapply(fit_algs, function(fit_alg) sag_fit(X, y,
+                                                    lambda=lambda,
+                                                    maxiter=maxiter,
+                                                    family=family,
+                                                    fit_alg=fit_alg,
+                                                    standardize=FALSE,
+                                                    tol=tol, monitor=TRUE))

> print(lapply(sag_fits, function(sag_fit) get_cost(sag_fit, X, y)))
Warning message:
In sag_fit(X, y, lambda = lambda, maxiter = maxiter, family = family,  :
  Optimisation stopped before convergence. Try incrasing maximum number of iterations
Error in print(lapply(sag_fits, function(sag_fit) get_cost(sag_fit, X,  : 
  error in evaluating the argument 'x' in selecting a method for function 'print': Error in FUN(X[[i]], ...) : could not find function "get_cost"
> sessionInfo()
R version 3.2.1 (2015-06-18)
Platform: x86_64-unknown-linux-gnu (64-bit)
Running under: Ubuntu precise (12.04.5 LTS)

locale:
 [1] LC_CTYPE=en_CA.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_CA.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_CA.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] glmnet_1.9-5        ggplot2_1.0.1       bigoptim_0.0.0.9000
[4] Matrix_1.2-1       

loaded via a namespace (and not attached):
 [1] Rcpp_0.11.6        lattice_0.20-31    digest_0.6.4       MASS_7.3-40       
 [5] grid_3.2.1         plyr_1.8.1         gtable_0.1.2       scales_0.2.3      
 [9] reshape2_1.2.2     labeling_0.2       proto_1.0.0        RColorBrewer_1.0-5
[13] tools_3.2.1        stringr_0.6.2      dichromat_2.0-0    munsell_0.4.2     
[17] colorspace_1.2-4  
> 

[IshmaelBelghazi]

Thanks for opening the issues. I have exported the root of the generic methods. Is it working now?

[tdhock]

not working, same error.

to debug on your own machine, try starting R from the command line with the
--vanilla option like I did in the code above.

On Mon, Aug 31, 2015 at 8:13 PM, Ishmael Belghazi notifications@github.com
wrote:

Thanks for opening the issues. I have exported the root of the generic
methods. Is it working now?

—
Reply to this email directly or view it on GitHub
#16 (comment)
.

[IshmaelBelghazi]

Thanks. Updated NAMESPACE. How about now?

[tdhock]

more problems. your results on the README look good though. are the timings for one lambda/regularization parameter, or for several?

thocking@silene:~/projects/blueprint(master)$ R --vanilla

R version 3.2.2 (2015-08-14) -- "Fire Safety"
Copyright (C) 2015 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

  Natural language support but running in an English locale

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

> 
> devtools::install_github("IshmaelBelghazi/bigoptim")
Installing github repo bigoptim/master from IshmaelBelghazi
Downloading master.zip from https://github.com/IshmaelBelghazi/bigoptim/archive/master.zip
Installing package from /tmp/Rtmp2apyUe/master.zip
arguments 'minimized' and 'invisible' are for Windows only
Installing bigoptim
'/home/thocking/lib/R/bin/R' --vanilla CMD INSTALL  \
  '/tmp/Rtmp2apyUe/devtools3ee078dc0a4d/bigoptim-master'  \
  --library='/home/thocking/lib/R/library' --install-tests 

* installing *source* package 'bigoptim' ...
** libs
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c dataset.c -o dataset.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c entrypoint-glm.c -o entrypoint-glm.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c entrypoint-sag.c -o entrypoint-sag.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c entrypoint-sag_fit.c -o entrypoint-sag_fit.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c glm_models.c -o glm_models.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_C_dynload_posix.c -o sag_C_dynload_posix.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_R_losses.c -o sag_R_losses.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_adaptive.c -o sag_adaptive.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_common.c -o sag_common.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_constant.c -o sag_constant.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_init.c -o sag_init.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_linesearch.c -o sag_linesearch.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_train.c -o sag_train.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c sag_warm.c -o sag_warm.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c trainers.c -o trainers.o
gcc -std=gnu99 -I/home/thocking/lib/R/include -DNDEBUG  -I/usr/local/include -I"/home/thocking/lib/R/library/Matrix/include"   -fpic  -g -O2  -c utils.c -o utils.o
gcc -std=gnu99 -shared -L/usr/local/lib -o bigoptim.so dataset.o entrypoint-glm.o entrypoint-sag.o entrypoint-sag_fit.o glm_models.o sag_C_dynload_posix.o sag_R_losses.o sag_adaptive.o sag_common.o sag_constant.o sag_init.o sag_linesearch.o sag_train.o sag_warm.o trainers.o utils.o -L/home/thocking/lib/R/lib -lRlapack -L/home/thocking/lib/R/lib -lRblas -lgfortran -lm -lquadmath
installing to /home/thocking/lib/R/library/bigoptim/libs
** R
** data
*** moving datasets to lazyload DB
** demo
** tests
** preparing package for lazy loading
** help
*** installing help indices
** building package indices
** testing if installed package can be loaded
* DONE (bigoptim)
> demo("monitoring_training")
Error in demo("monitoring_training") : 
  No demo found for topic ‘monitoring_training’
> library(bigoptim)
Loading required package: Matrix
^Pdemo("monitoring_training")


    demo(monitoring_training)
    ---- ~~~~~~~~~~~~~~~~~~~

Type  <Return>   to start : 

> suppressPackageStartupMessages(library(ggplot2))

> suppressPackageStartupMessages(library(glmnet))

> ## Loading Data set
> data(covtype.libsvm)

> ## Normalizing Columns and adding intercept
> X <- cbind(rep(1, NROW(covtype.libsvm$X)), scale(covtype.libsvm$X))

> y <- covtype.libsvm$y

> y[y == 2] <- -1

> n <- NROW(X)

> p <- NCOL(X)

> ## Setting seed
> #set.seed(0)
> ## Setting up problem
> n_passes <- 50  ## number of passses trough the dataset

> maxiter <- n * n_passes

> lambda <- 1/n 

> tol <- 0

> family <- "binomial"

> fit_algs <- list(constant="constant",
+                  linesearch="linesearch",
+                  adaptive="adaptive")

> sag_fits <- lapply(fit_algs, function(fit_alg) sag_fit(X, y,
+                                                    lambda=lambda,
+                                                    maxiter=maxiter,
+                                                    family=family,
+                                                    fit_alg=fit_alg,
+                                                    standardize=FALSE,
+                                                    tol=tol, monitor=TRUE))

> print(lapply(sag_fits, function(sag_fit) get_cost(sag_fit, X, y)))
$constant
[1] 0.5135133

$linesearch
[1] 0.5134914

$adaptive
[1] 0.521179


> ## Functions --------------------------------------------------------------------
> make_monitor_table <- function(object, X, y, omit_init_state=TRUE) {
+   monitor_fun <- list(cost=function(w)
+                              .get_cost(X, y, w,
+                                        object$input$lambda,
+                                        family=family,
+                                        backend="C"),
+                       grad_norm=function(w)
+                                   norm(.get_grad(X, y, w,
+                                                  lambda=object$input$lambda,
+                                                  family=family,
+                                                  backend="C"),
+                                        type='F'))
+   monitor_w <- object$monitor_w
+   if (omit_init_state) monitor_w <- monitor_w[, -1, drop=FALSE]
+   
+   mon <- lapply(monitor_fun, function(fun) apply(monitor_w, 2, fun))
+   mon$effective_pass <- seq(0, length(mon$cost) - 1)
+   if (omit_init_state) mon$effective_pass <- mon$effective_pass + 1
+   
+   mon$fit_alg <- rep(object$input$fit_alg, length(mon$cost))
+   as.data.frame(mon)
+ }

> ## Glmnet training table
> make_glmnet_monitor_table <- function(X, y, lambda, n_passes) {
+ 
+   glmnet_fit <- glmnet(X, as.factor(y), alpha=0, family="binomial",
+                        lambda=lambda, standardize=FALSE, intercept=FALSE)
+   glmnet_w <- as.matrix(coef(glmnet_fit))[-1]
+   glm_cost <- .get_cost(X, y, glmnet_w, lambda=lambda,
+                         family=family, backend="C")
+   glm_cost_grad <- norm(.get_grad(X, y, glmnet_w, lambda=lambda,
+                                   family=family, backend="C"), 'F')
+   effective_pass <- 1:n_passes
+   data.frame(cost=rep(glm_cost, n_passes),
+              grad_norm=rep(glm_cost_grad, n_passes),
+              effective_pass=effective_pass,
+              fit_alg="glmmnet")}

> ## Training monitoring functions
> ## Make graphs
> make_graph_by_pass <- function(training_table) {
+ 
+   training_table$fit_alg <- as.factor(training_table$fit_alg)
+   ## Cost monitoring graph
+   cost_graph <- ggplot(training_table,
+                        aes(x=effective_pass,
+                            y=cost,
+                            color=fit_alg)) +
+     geom_line() +
+     xlab("Number of effective passes") +
+     ylab("Cost") +
+     ggtitle("Cost per effective pass through the covtype dataset") +
+     scale_colour_discrete(name="Fit algorithm")
+ 
+   ## Gradient Frobenius norm monitoring graph
+   grad_norm_graph <- ggplot(training_table,
+                        aes(x=effective_pass,
+                            y=grad_norm,
+                            color=fit_alg)) +
+     geom_line() +
+     scale_y_log10() +
+     xlab("Number of effective passes") +
+     ylab("gradient L2 norm (Log scale)") +
+     ggtitle("Gradient L2 norm per effective pass through the covtype dataset") +
+     scale_colour_discrete(name="Fit algorithm")
+ 
+ 
+   list(cost=cost_graph, grad_norm=grad_norm_graph)
+ }

> ## Monitoring ------------------------------------------------------------------
> training_tables <- lapply(sag_fits, function(sag_fit)
+                                       make_monitor_table(sag_fit,
+                                                          X, y, 
+                                                          omit_init_state=TRUE))
Error in FUN(newX[, i], ...) : could not find function ".get_cost"
In addition: Warning message:
In sag_fit(X, y, lambda = lambda, maxiter = maxiter, family = family,  :
  Optimisation stopped before convergence. Try incrasing maximum number of iterations
> demo("run_times")


    demo(run_times)
    ---- ~~~~~~~~~

Type  <Return>   to start : 

> suppressPackageStartupMessages(library(ggplot2))

> suppressPackageStartupMessages(library(glmnet))

> family <- "binomial"

> fit_algs <- list(constant="constant",
+                  linesearch="linesearch",
+                  adaptive="adaptive")

> time_fit <- function(X, y, lambda, maxiter, family, tol){
+   fits <- lapply(fit_algs, function(fit_alg) {
+     print(sprintf("processing: %s", fit_alg))
+     fit_time <- system.time({fit <- sag_fit(X, y,
+                                             lambda=lambda,
+                                             maxiter=maxiter,
+                                             family=family,
+                                             fit_alg=fit_alg,
+                                             standardize=FALSE,
+                                             tol=tol)})
+     list(cost=get_cost(fit, X, y),
+          grad_norm=norm(get_grad(fit, X, y), 'F'),
+          approx_grad_norm=norm(get_approx_grad(fit), 'F'),
+          time=fit_time[['elapsed']])
+   })
+   print("processing: glmnet")
+   glmnet_time <- system.time({
+     glmnet_fit <- glmnet(X, as.factor(y), alpha=0, family="binomial",
+                          lambda=lambda, standardize=FALSE, intercept=FALSE)
+   })
+   if (is.sparse(X)) {
+     backend <- "R"
+     glmnet_w <- tail(coef(glmnet_fit), -1)
+   } else {
+     backend <- "C"
+     glmnet_w <- as.matrix(coef(glmnet_fit))[-1]
+   } 
+   glmnet_w <- as.matrix(coef(glmnet_fit))[-1]
+   backend <- if (is.sparse(X)) "R" else "C"
+   glm_cost <- .get_cost(X, y, glmnet_w, lambda=lambda,
+                         family=family, backend=backend)
+   glm_cost_grad <- norm(.get_grad(X, y, glmnet_w, lambda=lambda,
+                                   family=family, backend=backend), 'F')
+   fits$glmnet <- list(cost=glm_cost,
+                       grad_norm=glm_cost_grad,
+                       approx_grad_norm=NA,
+                       time=glmnet_time[['elapsed']])
+   do.call(cbind, fits)
+ }

> ## Covtype libsvm --------------------------------------------------------------
> ## Loading Data set
> data(covtype.libsvm)

> ## Normalizing Columns and adding intercept
> X_covtype <- cbind(rep(1, NROW(covtype.libsvm$X)), scale(covtype.libsvm$X))

> y_covtype <- covtype.libsvm$y

> y_covtype[y_covtype == 2] <- -1

> ## Setting seed
> maxiter <- NROW(X_covtype) * 10 

> lambda <- 1/NROW(X_covtype)

> tol <- 1e-4

> print("Timing on covertype ...")
[1] "Timing on covertype ..."

> fit_times_covtype <- time_fit(X_covtype, y_covtype, lambda, maxiter, family, tol)
[1] "processing: constant"
[1] "processing: linesearch"
[1] "processing: adaptive"
[1] "processing: glmnet"
Error in time_fit(X_covtype, y_covtype, lambda, maxiter, family, tol) : 
  could not find function "is.sparse"
In addition: Warning messages:
1: In sag_fit(X, y, lambda = lambda, maxiter = maxiter, family = family,  :
  Optimisation stopped before convergence. Try incrasing maximum number of iterations
2: In sag_fit(X, y, lambda = lambda, maxiter = maxiter, family = family,  :
  Optimisation stopped before convergence. Try incrasing maximum number of iterations
>