fix epoch and batch naming

R-package/NAMESPACE

@@ -40,12 +40,15 @@ export(mx.model.FeedForward.create)
 export(mx.model.load)
 export(mx.model.save)
 export(mx.nd.array)
+export(mx.nd.choose.element)
 export(mx.nd.clip)
 export(mx.nd.copyto)
 export(mx.nd.dot)
 export(mx.nd.load)
 export(mx.nd.ones)
 export(mx.nd.save)
+export(mx.nd.sqrt)
+export(mx.nd.square)
 export(mx.nd.zeros)
 export(mx.opt.create)
 export(mx.opt.get.updater)

R-package/R/model.R

@@ -96,8 +96,8 @@ mx.model.train <- function(symbol, ctx, input.shape,
                            begin.round, end.round, optimizer,
                            train.data, eval.data,
                            metric,
-                           iter.end.callback,
                            epoch.end.callback,
+                           batch.end.callback,
                            kvstore) {
   ndevice <- length(ctx)
   cat(paste0("Start training with ", ndevice, " devices\n"))
@@ -195,8 +195,8 @@ mx.model.train <- function(symbol, ctx, input.shape,
         }
       }
       nbatch <- nbatch + 1
-      if (!is.null(epoch.end.callback)) {
-        epoch.end.callback(iteration, nbatch, environment())
+      if (!is.null(batch.end.callback)) {
+        batch.end.callback(iteration, nbatch, environment())
       }
     }
     # reset training data
@@ -244,8 +244,8 @@ mx.model.train <- function(symbol, ctx, input.shape,
     }
     # get the model out
     model <- mx.model.extract.model(symbol, train.execs)
-    if (!is.null(iter.end.callback)) {
-      iter.end.callback(iteration, 0, environment())
+    if (!is.null(epoch.end.callback)) {
+      epoch.end.callback(iteration, 0, environment())
     }
   }
   return(model)
@@ -355,9 +355,9 @@ mx.model.select.layout.predict <- function(X, model) {
 #'     The validation set used for validation evaluation during the progress
 #' @param eval.metric function, optional
 #'     The evaluation function on the results.
-#' @param iter.end.callback function, optional
-#'     The callback when iteration ends.
 #' @param epoch.end.callback function, optional
+#'     The callback when iteration ends.
+#' @param batch.end.callback function, optional
 #'     The callback when one mini-batch iteration ends.
 #' @param array.batch.size integer (default=128)
 #'     The batch size used for R array training.
@@ -377,7 +377,7 @@ function(symbol, X, y=NULL, ctx=NULL,
          num.round=10, optimizer="sgd",
          initializer=mx.init.uniform(0.01),
          eval.data=NULL, eval.metric=NULL,
-         iter.end.callback=NULL, epoch.end.callback=NULL,
+         epoch.end.callback=NULL, batch.end.callback=NULL,
          array.batch.size=128, array.layout="auto",
          kvstore="local",
          ...) {
@@ -413,8 +413,8 @@ function(symbol, X, y=NULL, ctx=NULL,
                           1, num.round, optimizer=optimizer,
                           train.data=X, eval.data=eval.data,
                           metric=eval.metric,
-                          iter.end.callback=iter.end.callback,
                           epoch.end.callback=epoch.end.callback,
+                          batch.end.callback=batch.end.callback,
                           kvstore=kvstore)
   return (model)
 }

R-package/demo/basic_convnet.R

@@ -47,5 +47,5 @@ model <- mx.model.FeedForward.create(softmax, X=dtrain, eval.data=dtest,
                                      ctx=devices, num.round=1,
                                      learning.rate=0.1, momentum=0.9,
                                      initializer=mx.init.uniform(0.07),
-                                     epoch.end.callback=mx.callback.log.train.metric(100))
+                                     batch.end.callback=mx.callback.log.train.metric(100))
 

R-package/demo/basic_model.R

@@ -39,8 +39,8 @@ model <- mx.model.FeedForward.create(softmax, X=dtrain, eval.data=dtest,
                                      ctx=devices, num.round=1,
                                      learning.rate=0.1, momentum=0.9,
                                      initializer=mx.init.uniform(0.07),
-                                     iter.end.callback=mx.callback.save.checkpoint("chkpt"),
-                                     epoch.end.callback=mx.callback.log.train.metric(100))
+                                     epoch.end.callback=mx.callback.save.checkpoint("chkpt"),
+                                     batch.end.callback=mx.callback.log.train.metric(100))
 
 # do prediction
 pred <- predict(model, dtest)

R-package/demo/basic_training.R

@@ -41,4 +41,4 @@ model <- mx.model.FeedForward.create(softmax, X=X, y=y,
                                      ctx=devices, num.round=1,
                                      learning.rate=0.1, momentum=0.9,
                                      initializer=mx.init.uniform(0.07),
-                                     epoch.end.callback=mx.callback.log.train.metric(100))
+                                     batch.end.callback=mx.callback.log.train.metric(100))

R-package/man/mx.model.FeedForward.create.Rd

@@ -6,9 +6,9 @@
 \usage{
 mx.model.FeedForward.create(symbol, X, y = NULL, ctx = NULL,
   num.round = 10, optimizer = "sgd", initializer = mx.init.uniform(0.01),
-  eval.data = NULL, eval.metric = NULL, iter.end.callback = NULL,
-  epoch.end.callback = NULL, array.batch.size = 128, kvstore = "local",
-  ...)
+  eval.data = NULL, eval.metric = NULL, epoch.end.callback = NULL,
+  batch.end.callback = NULL, array.batch.size = 128,
+  array.layout = "auto", kvstore = "local", ...)
 }
 \arguments{
 \item{symbol}{The symbolic configuration of the neural network.}
@@ -37,15 +37,22 @@ The validation set used for validation evaluation during the progress}
 \item{eval.metric}{function, optional
 The evaluation function on the results.}
 
-\item{iter.end.callback}{function, optional
+\item{epoch.end.callback}{function, optional
 The callback when iteration ends.}
 
-\item{epoch.end.callback}{function, optional
+\item{batch.end.callback}{function, optional
 The callback when one mini-batch iteration ends.}
 
 \item{array.batch.size}{integer (default=128)
 The batch size used for R array training.}
 
+\item{array.layout}{can be "auto", "colmajor", "rowmajor", (detault=auto)
+The layout of array. "rowmajor" is only supported for two dimensional array.
+For matrix, "rowmajor" means dim(X) = c(nexample, nfeatures),
+"colmajor" means dim(X) = c(nfeatures, nexample)
+"auto" will auto detect the layout by match the feature size,
+ and will report error when X is a square matrix to ask user to explicitly specify layout.}
+
 \item{kvstore}{string (default="local")
 The parameter synchronization scheme in multiple devices.}
 }

R-package/man/mx.nd.choose.element.Rd

@@ -0,0 +1,19 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mxnet_generated.R
+\name{mx.nd.choose.element}
+\alias{mx.nd.choose.element}
+\title{Choose one element from each line(row for python, column for R/Julia) in lhs according to index indicated by rhs}
+\arguments{
+\item{lhs}{NDArray
+Left operand to the function.}
+
+\item{rhs}{NDArray
+Right operand to the function.}
+}
+\value{
+out The result mx.ndarray
+}
+\description{
+Choose one element from each line(row for python, column for R/Julia) in lhs according to index indicated by rhs
+}
+

R-package/man/mx.nd.sqrt.Rd

@@ -0,0 +1,16 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mxnet_generated.R
+\name{mx.nd.sqrt}
+\alias{mx.nd.sqrt}
+\title{Take square root of the src}
+\arguments{
+\item{src}{NDArray
+Source input to the function.}
+}
+\value{
+out The result mx.ndarray
+}
+\description{
+Take square root of the src
+}
+

R-package/man/mx.nd.square.Rd

@@ -0,0 +1,16 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mxnet_generated.R
+\name{mx.nd.square}
+\alias{mx.nd.square}
+\title{Take square of the src}
+\arguments{
+\item{src}{NDArray
+Source input to the function.}
+}
+\value{
+out The result mx.ndarray
+}
+\description{
+Take square of the src
+}
+

R-package/man/predict.MXFeedForwardModel.Rd

@@ -5,7 +5,7 @@
 \title{Predict the outputs given a model and dataset.}
 \usage{
 \method{predict}{MXFeedForwardModel}(model, X, ctx = NULL,
-  array.batch.size = 128)
+  array.batch.size = 128, array.layout = "auto")
 }
 \arguments{
 \item{model}{The MXNet Model.}
@@ -15,6 +15,13 @@
 \item{ctx}{mx.cpu() or mx.gpu(i) The device used to generate the prediction.}
 
 \item{array.batch.size}{The batch size used in batching. Only used when X is R's array.}
+
+\item{array.layout}{can be "auto", "colmajor", "rowmajor", (detault=auto)
+    The layout of array. "rowmajor" is only supported for two dimensional array.
+    For matrix, "rowmajor" means dim(X) = c(nexample, nfeatures),
+    "colmajor" means dim(X) = c(nfeatures, nexample)
+    "auto" will auto detect the layout by match the feature size,
+     and will report error when X is a square matrix to ask user to explicitly specify layout.}
 }
 \description{
 Predict the outputs given a model and dataset.