Skip to content
This repository was archived by the owner on Nov 17, 2023. It is now read-only.

[R] Fix current docs #322

Merged
tqchen merged 2 commits into from
Oct 18, 2015
Merged

[R] Fix current docs #322

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
ab62f70
Merge pull request #6 from dmlc/master
hetong007 Oct 18, 2015
b9363ba
fix docs
hetong007 Oct 18, 2015
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 22 additions & 11 deletions R-package/R/model.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -133,7 +133,9 @@ mx.model.train <- function(symbol, ctx, input.shape,

for (iteration in begin.round:end.round) {
nbatch <- 0
train.metric <- metric$init()
if (!is.null(metric)) {
train.metric <- metric$init()
}
while (train.data$iter.next()) {
# Get input data slice
dlist <- train.data$value()
Expand Down Expand Up @@ -186,8 +188,10 @@ mx.model.train <- function(symbol, ctx, input.shape,
}
}
# Update the evaluation metrics
for (i in 1 : ndevice) {
train.metric <- metric$update(slices[[i]]$label, out.preds[[i]], train.metric)
if (!is.null(metric)) {
for (i in 1 : ndevice) {
train.metric <- metric$update(slices[[i]]$label, out.preds[[i]], train.metric)
}
}
nbatch <- nbatch + 1
if (!is.null(epoch.end.callback)) {
Expand All @@ -196,11 +200,14 @@ mx.model.train <- function(symbol, ctx, input.shape,
}
# reset training data
train.data$reset()
result <- metric$get(train.metric)
cat(paste0("[", iteration, "] Train-", result$name, "=", result$value, "\n"))

if (!is.null(metric)) {
result <- metric$get(train.metric)
cat(paste0("[", iteration, "] Train-", result$name, "=", result$value, "\n"))
}
if (!is.null(eval.data)) {
eval.metric <- metric$init()
if (!is.null(metric)) {
eval.metric <- metric$init()
}
while (eval.data$iter.next()) {
dlist <- eval.data$value()
slices <- lapply(1:ndevice, function(i) {
Expand All @@ -220,13 +227,17 @@ mx.model.train <- function(symbol, ctx, input.shape,
out.preds <- lapply(train.execs, function(texec) {
mx.nd.copyto(texec$ref.outputs[[1]], mx.cpu())
})
for (i in 1 : ndevice) {
eval.metric <- metric$update(slices[[i]]$label, out.preds[[i]], eval.metric)
if (!is.null(metric)) {
for (i in 1 : ndevice) {
eval.metric <- metric$update(slices[[i]]$label, out.preds[[i]], eval.metric)
}
}
}
eval.data$reset()
result <- metric$get(eval.metric)
cat(paste0("[", iteration, "] Validation-", result$name, "=", result$value, "\n"))
if (!is.null(metric)) {
result <- metric$get(eval.metric)
cat(paste0("[", iteration, "] Validation-", result$name, "=", result$value, "\n"))
}
} else {
eval.metric <- NULL
}
Expand Down
38 changes: 17 additions & 21 deletions R-package/vignettes/fiveMinutesNeuralNetwork.Rmd
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,12 @@ This is the first tutorial for new users of the R package `mxnet`. You will lear

We will show you how to do classification and regression tasks respectively. The data we use comes from the package `mlbench`.

Preface
-------
This tutorial is written in Rmarkdown.
- You can directly view the hosted version of the tutorial from [MXNet R Document](http://mxnet.readthedocs.org/en/latest/R-package/fiveMinutesNeuralNetwork.html)
- You can find the download the Rmarkdown source from [here](https://github.com/dmlc/mxnet/blob/master/R-package/vignettes/fiveMinutesNeuralNetwork.Rmd)

## Classification

First of all, let us load in the data and preprocess it:
Expand All @@ -30,19 +36,17 @@ The next step is to define the structure of the neural network.
data <- mx.symbol.Variable("data")
# A fully connected hidden layer
# data: input source
# name: fc1
# num_hidden: number of neurons in this hidden layer
fc1 <- mx.symbol.FullyConnected(data, name="fc1", num_hidden=20)
fc1 <- mx.symbol.FullyConnected(data, num_hidden=20)

# An activation function
# fc1: input source
# name: relu1
# act_type: type for the activation function
act1 <- mx.symbol.Activation(fc1, name="tanh1", act_type="tanh")
fc2 <- mx.symbol.FullyConnected(act1, name="fc2", num_hidden=2)
act1 <- mx.symbol.Activation(fc1, act_type="tanh")
fc2 <- mx.symbol.FullyConnected(act1, num_hidden=2)

# Softmax function for the output layer
softmax <- mx.symbol.Softmax(fc2, name="sm")
softmax <- mx.symbol.Softmax(fc2)
```

According to the comments in the code, you can see the meaning of each function and its arguments. They can be easily modified according to your need.
Expand Down Expand Up @@ -87,35 +91,27 @@ test.x = data.matrix(BostonHousing[-train.ind, -14])
test.y = BostonHousing[-train.ind, 14]
```

We can configure a similar network as what we have done above. The only difference is in the output activation:
We can configure another network as what we have done above. The main difference is in the output activation:

```{r}
# Define the input data
data <- mx.symbol.Variable("data")
# A fully connected hidden layer
# data: input source
# name: fc1
# num_hidden: number of neurons in this hidden layer
fc1 <- mx.symbol.FullyConnected(data, name="fc1", num_hidden=20)

# An activation function
# fc1: input source
# name: relu1
# act_type: type for the activation function
act1 <- mx.symbol.Activation(fc1, name="tanh1", act_type="tanh")
fc2 <- mx.symbol.FullyConnected(act1, name="fc2", num_hidden=1)
fc1 <- mx.symbol.FullyConnected(data, num_hidden=1)

# Softmax function for the output layer
lro <- mx.symbol.LinearRegressionOutput(fc2, name="lro")
lro <- mx.symbol.LinearRegressionOutput(fc1)
```

What we changed is mainly the last function, this enables the new network to optimize for squared loss. We can now train on this simple data set.

```{r}
mx.set.seed(0)
model <- mx.model.FeedForward.create(lro, X=train.x, y=train.y,
ctx=device.cpu, num.round=5, array.batch.size=10,
learning.rate=0.1, momentum=0.9, eval.metric=mx.metric.rmse,
ctx=device.cpu, num.round=50, array.batch.size=20,
learning.rate=2e-6, momentum=0.9, eval.metric=mx.metric.rmse,
epoch.end.callback=mx.callback.log.train.metric(100))
```

Expand All @@ -140,8 +136,8 @@ This is an example for mean absolute error. We can simply plug it in the trainin
```{r}
mx.set.seed(0)
model <- mx.model.FeedForward.create(lro, X=train.x, y=train.y,
ctx=device.cpu, num.round=5, array.batch.size=10,
learning.rate=0.1, momentum=0.9, eval.metric=demo.metric.mae,
ctx=device.cpu, num.round=50, array.batch.size=20,
learning.rate=2e-6, momentum=0.9, eval.metric=demo.metric.mae,
epoch.end.callback=mx.callback.log.train.metric(100))
```

Expand Down
150 changes: 118 additions & 32 deletions doc/R-package/fiveMinutesNeuralNetwork.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,12 @@ This is the first tutorial for new users of the R package `mxnet`. You will lear

We will show you how to do classification and regression tasks respectively. The data we use comes from the package `mlbench`.

Preface
-------
This tutorial is written in Rmarkdown.
- You can directly view the hosted version of the tutorial from [MXNet R Document](http://mxnet.readthedocs.org/en/latest/R-package/fiveMinutesNeuralNetwork.html)
- You can find the download the Rmarkdown source from [here](https://github.com/dmlc/mxnet/blob/master/R-package/vignettes/fiveMinutesNeuralNetwork.Rmd)

## Classification

First of all, let us load in the data and preprocess it:
Expand Down Expand Up @@ -46,19 +52,17 @@ The next step is to define the structure of the neural network.
data <- mx.symbol.Variable("data")
# A fully connected hidden layer
# data: input source
# name: fc1
# num_hidden: number of neurons in this hidden layer
fc1 <- mx.symbol.FullyConnected(data, name="fc1", num_hidden=20)
fc1 <- mx.symbol.FullyConnected(data, num_hidden=20)

# An activation function
# fc1: input source
# name: relu1
# act_type: type for the activation function
act1 <- mx.symbol.Activation(fc1, name="tanh1", act_type="tanh")
fc2 <- mx.symbol.FullyConnected(act1, name="fc2", num_hidden=2)
act1 <- mx.symbol.Activation(fc1, act_type="tanh")
fc2 <- mx.symbol.FullyConnected(act1, num_hidden=2)

# Softmax function for the output layer
softmax <- mx.symbol.Softmax(fc2, name="sm")
softmax <- mx.symbol.Softmax(fc2)
```

According to the comments in the code, you can see the meaning of each function and its arguments. They can be easily modified according to your need.
Expand Down Expand Up @@ -138,27 +142,19 @@ test.x = data.matrix(BostonHousing[-train.ind, -14])
test.y = BostonHousing[-train.ind, 14]
```

We can configure a similar network as what we have done above. The only difference is in the output activation:
We can configure another network as what we have done above. The main difference is in the output activation:


```r
# Define the input data
data <- mx.symbol.Variable("data")
# A fully connected hidden layer
# data: input source
# name: fc1
# num_hidden: number of neurons in this hidden layer
fc1 <- mx.symbol.FullyConnected(data, name="fc1", num_hidden=20)

# An activation function
# fc1: input source
# name: relu1
# act_type: type for the activation function
act1 <- mx.symbol.Activation(fc1, name="tanh1", act_type="tanh")
fc2 <- mx.symbol.FullyConnected(act1, name="fc2", num_hidden=1)
fc1 <- mx.symbol.FullyConnected(data, num_hidden=1)

# Softmax function for the output layer
lro <- mx.symbol.LinearRegressionOutput(fc2, name="lro")
lro <- mx.symbol.LinearRegressionOutput(fc1)
```

What we changed is mainly the last function, this enables the new network to optimize for squared loss. We can now train on this simple data set.
Expand All @@ -167,18 +163,63 @@ What we changed is mainly the last function, this enables the new network to opt
```r
mx.set.seed(0)
model <- mx.model.FeedForward.create(lro, X=train.x, y=train.y,
ctx=device.cpu, num.round=5, array.batch.size=10,
learning.rate=0.1, momentum=0.9, eval.metric=mx.metric.rmse,
ctx=device.cpu, num.round=50, array.batch.size=20,
learning.rate=2e-6, momentum=0.9, eval.metric=mx.metric.rmse,
epoch.end.callback=mx.callback.log.train.metric(100))
```

```
## Start training with 1 devices
## [1] Train-rmse=20.8877275599495
## [2] Train-rmse=12.8786644532322
## [3] Train-rmse=10.3635559222185
## [4] Train-rmse=10.5605206622052
## [5] Train-rmse=10.2502398389275
## [1] Train-rmse=16.063282524034
## [2] Train-rmse=12.2792375712573
## [3] Train-rmse=11.1984634005885
## [4] Train-rmse=10.2645236892904
## [5] Train-rmse=9.49711005504284
## [6] Train-rmse=9.07733734175182
## [7] Train-rmse=9.07884450847991
## [8] Train-rmse=9.10463850277417
## [9] Train-rmse=9.03977049028532
## [10] Train-rmse=8.96870685004475
## [11] Train-rmse=8.93113287361574
## [12] Train-rmse=8.89937257821847
## [13] Train-rmse=8.87182096922953
## [14] Train-rmse=8.84476075083586
## [15] Train-rmse=8.81464673014974
## [16] Train-rmse=8.78672567900196
## [17] Train-rmse=8.76265872846474
## [18] Train-rmse=8.73946101419974
## [19] Train-rmse=8.71651926303267
## [20] Train-rmse=8.69457600919277
## [21] Train-rmse=8.67354928674563
## [22] Train-rmse=8.65328755392436
## [23] Train-rmse=8.63378039680078
## [24] Train-rmse=8.61488162586984
## [25] Train-rmse=8.5965105183022
## [26] Train-rmse=8.57868133563275
## [27] Train-rmse=8.56135851937663
## [28] Train-rmse=8.5444819772098
## [29] Train-rmse=8.52802114610432
## [30] Train-rmse=8.5119504512622
## [31] Train-rmse=8.49624261719241
## [32] Train-rmse=8.48087453238701
## [33] Train-rmse=8.46582689119887
## [34] Train-rmse=8.45107881002491
## [35] Train-rmse=8.43661331401712
## [36] Train-rmse=8.42241575909639
## [37] Train-rmse=8.40847217331365
## [38] Train-rmse=8.39476931796395
## [39] Train-rmse=8.38129658373974
## [40] Train-rmse=8.36804269059018
## [41] Train-rmse=8.35499817678397
## [42] Train-rmse=8.34215505742154
## [43] Train-rmse=8.32950441908131
## [44] Train-rmse=8.31703985777311
## [45] Train-rmse=8.30475363906755
## [46] Train-rmse=8.29264031506106
## [47] Train-rmse=8.28069372820073
## [48] Train-rmse=8.26890902770415
## [49] Train-rmse=8.25728089053853
## [50] Train-rmse=8.24580511500735
```

It is also easy to make prediction and evaluate
Expand All @@ -190,7 +231,7 @@ sqrt(mean((preds-test.y)^2))
```

```
## [1] 9.49181
## [1] 7.800502
```

Currently we have two pre-defined metrics "accuracy" and "rmse". One might wonder how to customize the evaluation metric. `mxnet` provides the interface for users to define their own metric of interests:
Expand All @@ -209,18 +250,63 @@ This is an example for mean absolute error. We can simply plug it in the trainin
```r
mx.set.seed(0)
model <- mx.model.FeedForward.create(lro, X=train.x, y=train.y,
ctx=device.cpu, num.round=5, array.batch.size=10,
learning.rate=0.1, momentum=0.9, eval.metric=demo.metric.mae,
ctx=device.cpu, num.round=50, array.batch.size=20,
learning.rate=2e-6, momentum=0.9, eval.metric=demo.metric.mae,
epoch.end.callback=mx.callback.log.train.metric(100))
```

```
## Start training with 1 devices
## [1] Train-mae=19.3546375619262
## [2] Train-mae=10.5938747770646
## [3] Train-mae=8.51244305161869
## [4] Train-mae=8.41277845326592
## [5] Train-mae=8.23570416674895
## [1] Train-mae=13.1889538083225
## [2] Train-mae=9.81431959337658
## [3] Train-mae=9.21576419870059
## [4] Train-mae=8.38071537613869
## [5] Train-mae=7.45462437611487
## [6] Train-mae=6.93423301743136
## [7] Train-mae=6.91432357016537
## [8] Train-mae=7.02742733055105
## [9] Train-mae=7.00618194618469
## [10] Train-mae=6.92541576984028
## [11] Train-mae=6.87530243690643
## [12] Train-mae=6.84757369098564
## [13] Train-mae=6.82966501611388
## [14] Train-mae=6.81151759574811
## [15] Train-mae=6.78394182841811
## [16] Train-mae=6.75914719419347
## [17] Train-mae=6.74180388773481
## [18] Train-mae=6.725853071279
## [19] Train-mae=6.70932178215848
## [20] Train-mae=6.6928868798746
## [21] Train-mae=6.6769521329138
## [22] Train-mae=6.66184809505939
## [23] Train-mae=6.64754504809777
## [24] Train-mae=6.63358514060577
## [25] Train-mae=6.62027640889088
## [26] Train-mae=6.60738245232238
## [27] Train-mae=6.59505546771818
## [28] Train-mae=6.58346195800437
## [29] Train-mae=6.57285477783945
## [30] Train-mae=6.56259003960424
## [31] Train-mae=6.5527790788975
## [32] Train-mae=6.54353428422991
## [33] Train-mae=6.5344172368447
## [34] Train-mae=6.52557652526432
## [35] Train-mae=6.51697905850079
## [36] Train-mae=6.50847898812758
## [37] Train-mae=6.50014844106303
## [38] Train-mae=6.49207674844397
## [39] Train-mae=6.48412070125341
## [40] Train-mae=6.47650500999557
## [41] Train-mae=6.46893867486053
## [42] Train-mae=6.46142131653097
## [43] Train-mae=6.45395035048326
## [44] Train-mae=6.44652914123403
## [45] Train-mae=6.43916216409869
## [46] Train-mae=6.43183777381976
## [47] Train-mae=6.42455544223388
## [48] Train-mae=6.41731406417158
## [49] Train-mae=6.41011292926139
## [50] Train-mae=6.40312503493494
```

Congratulations! Now you have learnt the basic for using `mxnet`.
Expand Down
6 changes: 3 additions & 3 deletions doc/R-package/index.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,5 @@
MXNet R Packge: Deep Learning for R
===================================

MXNet R Package
===============
You have find MXNet R Package! The MXNet R packages brings flexible and efficient GPU
computing and state-of-art deep learning to R.

Expand All @@ -21,6 +20,7 @@ There are several information to get you started

Tutorials
---------
* [Neural Network with MXNet in Five Minutes](fiveMinutesNeuralNetwork.md)
* [Classify Realworld Images with Pretrained Model](classifyRealImageWithPretrainedModel.md)
* [Handwritten Digits Classification Competition](mnistCompetition.md)
* [Tutorial on NDArray and Symbol](ndarrayAndSymbolTutorial.md)
Expand Down