ruby-dnn is a ruby deep learning library. This library supports full connected neural network and convolution neural network and recurrent neural network. Currently, you can get 99% accuracy with MNIST and 82% with CIFAR 10.
Add this line to your application's Gemfile:
gem 'ruby-dnn'
And then execute:
$ bundle
Or install it yourself as:
$ gem install ruby-dnn
model = Sequential.new
model << InputLayer.new(784)
model << Dense.new(256)
model << ReLU.new
model << Dense.new(256)
model << ReLU.new
model << Dense.new(10)
model.setup(Adam.new, SoftmaxCrossEntropy.new)
trainer = DNN::Trainer.new(model)
trainer.fit(x_train, y_train, 10, batch_size: 128, test: [x_test, y_test])
accuracy, loss = trainer.evaluate(x_test, y_test)
puts "accuracy: #{accuracy}"
puts "loss: #{loss}"
When create a model with 'define by run' style:
class MLP < Model
def initialize
super
@d1 = Dense.new(256)
@d2 = Dense.new(256)
@d3 = Dense.new(10)
end
def forward(x)
fs = DNN::Functions::FunctionSpace
x = InputLayer.new(784).(x)
x = @d1.(x)
x = fs.relu(x)
x = @d2.(x)
x = fs.relu(x)
x = @d3.(x)
x
end
end
model = MLP.new
model.setup(Adam.new, SoftmaxCrossEntropy.new)
trainer = DNN::Trainer.new(model)
trainer.fit(x_train, y_train, 10, batch_size: 128, test: [x_test, y_test])
accuracy, loss = trainer.evaluate(x_test, y_test)
puts "accuracy: #{accuracy}"
puts "loss: #{loss}"
Please refer to examples for basic usage.
If you want to know more detailed information, please refer to the source code.
-
Pix2pix
Convert an abstract image into a building image.
https://github.com/unagiootoro/facade-pix2pix -
Cycle-GAN
Convert apples to oranges and oranges to apples.
https://github.com/unagiootoro/apple2orange-cyclegan -
DQN
Learn the game so that the pole on the cart does not fall.
https://github.com/unagiootoro/ruby-rl
Implemented classes | |
---|---|
Connections | Dense, Conv2D, Conv2DTranspose, Embedding, SimpleRNN, LSTM, GRU |
Activations | Sigmoid, Tanh, Softsign, Softplus, Swish, ReLU, LeakyReLU, ELU, Mish |
Basic | Flatten, Reshape, Dropout, BatchNormalization |
Pooling | MaxPool2D, AvgPool2D, GlobalAvgPool2D, UnPool2D |
Optimizers | SGD, Nesterov, AdaGrad, RMSProp, AdaDelta, RMSPropGraves, Adam, AdaBound |
Losses | MeanSquaredError, MeanAbsoluteError, Hinge, HuberLoss, SoftmaxCrossEntropy, SigmoidCrossEntropy |
By setting the environment variable RUBY_DNN_DOWNLOADS_PATH
, you can specify the path to read dataset.
- Iris
- MNIST
- Fashion-MNIST
- CIFAR-10
- CIFAR-100
- STL-10
If you do require "cumo/narray"
before require "dnn"
, you can run it on GPU.
Or, set the environment variable RUBY_DNN_USE_CUMO
to ENABLE
to force the GPU to be used.
When running on a CPU, you can speed it up by using Numo Linalg.
In this case, Numo Linalg is automatically loaded by setting the environment variable RUNY_DNN_USE_NUMO_LINALG
to ENABLE
.
- Write a test.
- Write a document.
- Improve performance when using GPU.
After checking out the repo, run bin/setup
to install dependencies. Then, run rake "test"
to run the tests. You can also run bin/console
for an interactive prompt that will allow you to experiment.
To install this gem onto your local machine, run bundle exec rake install
. To release a new version, update the version number in version.rb
, and then run bundle exec rake release
, which will create a git tag for the version, push git commits and tags, and push the .gem
file to rubygems.org.
Bug reports and pull requests are welcome on GitHub at https://github.com/unagiootoro/ruby-dnn. This project is intended to be a safe, welcoming space for collaboration, and contributors are expected to adhere to the Contributor Covenant code of conduct.
The gem is available as open source under the terms of the MIT License.
Everyone interacting in the ruby-dnn project’s codebases, issue trackers, chat rooms and mailing lists is expected to follow the code of conduct.