TastyPancakes/Flux.jl
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

update logreg example

Browse Source
This commit is contained in:
Mike J Innes 2017-03-09 00:23:04 +00:00
parent 15b3ce1ada
commit 5c548fe1c4
3 changed files with 6 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/make.jl
View File

@ -17,7 +17,7 @@ makedocs(modules=[Flux],
"Backends" => "apis/backends.md", "Backends" => "apis/backends.md",
"Storing Models" => "apis/storage.md"], "Storing Models" => "apis/storage.md"],
"In Action" => [ "In Action" => [
"Logistic Regression" => "examples/logreg.md", "Simple MNIST" => "examples/logreg.md",
"Char RNN" => "examples/char-rnn.md"], "Char RNN" => "examples/char-rnn.md"],
"Contributing & Help" => "contributing.md", "Contributing & Help" => "contributing.md",
"Internals" => "internals.md"]) "Internals" => "internals.md"])

4
docs/src/examples/logreg.md
View File

@ -1,4 +1,4 @@
# Logistic Regression with MNIST # Recognising MNIST Digits
This walkthrough example will take you through writing a multi-layer perceptron that classifies MNIST digits with high accuracy. This walkthrough example will take you through writing a multi-layer perceptron that classifies MNIST digits with high accuracy.
@ -40,7 +40,7 @@ m = Chain(
Affine( 64), relu, Affine( 64), relu,
Affine( 10), softmax) Affine( 10), softmax)
model = tf(m) model = mxnet(m) # Convert to MXNet
``` ```
We can try this out on our data already: We can try this out on our data already:

6
examples/MNIST.jl
View File

@ -10,13 +10,13 @@ m = Chain(
Affine( 64), relu, Affine( 64), relu,
Affine( 10), softmax) Affine( 10), softmax)
# Convert to TensorFlow # Convert to MXNet
model = tf(m) model = mxnet(m)
# An example prediction pre-training # An example prediction pre-training
model(data[1][1]) model(data[1][1])
@time Flux.train!(model, train, test, η = 1e-4) Flux.train!(model, train, test, η = 1e-4)
# An example prediction post-training # An example prediction post-training
model(data[1][1]) model(data[1][1])