lstm

README

This is an LSTM implementation is pure go made with gorgonia.

The documentation is in progress.

By now, you can go get github.com/owulveryck/lstm and then run the example:

cd example/train ; go run ../../data/tontons/input.txt It will train the LSTM and predict every now and then.

TODO: the Gorgonia API has changed, I may need to asjust the solver when the 0.9 will be released

Documentation

Index

• type Model
  • func NewModel(inputSize, outputSize int, hiddenSize int) *Model
  • func (m Model) MarshalBinary() ([]byte, error)
  • func (m *Model) Predict(ctx context.Context, dataSet datasetter.Float32ReadWriter) error
  • func (m *Model) Train(ctx context.Context, dset datasetter.FullTrainer, solver G.Solver, ...) (<-chan TrainingInfos, <-chan error)
  • func (m *Model) UnmarshalBinary(data []byte) error
• type TrainingInfos

Types

type Model

type Model struct {
	// contains filtered or unexported fields
}

Model holds the tensor of the model

func NewModel

func NewModel(inputSize, outputSize int, hiddenSize int) *Model

NewModel creates a new model

func (Model) MarshalBinary

func (m Model) MarshalBinary() ([]byte, error)

MarshalBinary for backup. This function saves the content of the weights matrices and the biais but not the graph structure

func (*Model) Predict

func (m *Model) Predict(ctx context.Context, dataSet datasetter.Float32ReadWriter) error

Predict ...

func (*Model) Train

func (m *Model) Train(ctx context.Context, dset datasetter.FullTrainer, solver G.Solver, pauseChan <-chan struct{}) (<-chan TrainingInfos, <-chan error)

Train the model

func (*Model) UnmarshalBinary

func (m *Model) UnmarshalBinary(data []byte) error

UnmarshalBinary for restore

type TrainingInfos

type TrainingInfos struct {
	Step       int
	Perplexity float32
	Cost       float32
}

TrainingInfos returns info about the current training process