Documentation
¶
Index ¶
- func Configure() error
- type Char
- func (c *Char) ApplyDist(p []float64) []float64
- func (c *Char) Decode(xs [][]float64) io.Reader
- func (c *Char) Encode(r io.Reader) [][]float64
- func (c *Char) Feed() <-chan rnn.TrainingSet
- func (c *Char) GetInfos() json.Marshaler
- func (c *Char) MarshalBinary() ([]byte, error)
- func (c *Char) NewRNN() *rnn.RNN
- func (c *Char) SetLoss(loss float64)
- func (c *Char) UnmarshalBinary(b []byte) error
- type Infos
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
Types ¶
type Char ¶
type Char struct {
// contains filtered or unexported fields
}
Char is the basic codec for feeding a RNN with text
func (*Char) ApplyDist ¶
ApplyDist applies a distribution according to the configuration of the neural network
func (*Char) Decode ¶
Decode an array of inputs and returns an io.Reader the input is an array of 1-of-K encoded vectors
func (*Char) Feed ¶
func (c *Char) Feed() <-chan rnn.TrainingSet
Feed returns a channel that will be filled with TrainingSets its triggers a go-routine that reads the input and that is putting some data in the channel
Source Files
Click to show internal directories.
Click to hide internal directories.