mlp_train.c File Reference

Training of an MLP. More...

Functions
SprNNFloatData *	mlp_train_fwd (SprMLP *mlp, float *x)
double	mlp_train_bwd (SprMLP *mlp)
SprNNFloatCalc	mlp_train_err_ls (SprMLP *mlp, const float *t, const float *wgt, SprNNFloatCalc eps)
SprNNFloatCalc	mlp_train_err_ml (SprMLP *mlp, int Ndx, SprNNFloatCalc eps)
SprNNFloatCalc	mlp_train_err_ml1 (SprMLP *mlp, int Ndx, SprNNFloatCalc eps)
int	mlp_update (SprMLP *mlp, int final)
int	mlp_update_WGDI (SprMLP *mlp)
void	spr_mlp_train_stats (SprStream *fd, SprMLP *mlp)
int	spr_mlp_train_state_io (const char *fname, SprMLP *mlp, int save)

Detailed Description

Training of an MLP.

Forward evaluation (with intermediate results) and backward updating (error back propagation) of all connections in an MLP.

Note

The error functions (mlp_train_err_xxx) return the unweighted frame error, i.e. each frame is assumed to have a weight of 1.0. The gradients however can be weighted by varying the learning rate (eps). Making the weigthing average of the error must be done by the main program.

Date

Feb 1999

Author

Kris Demuynck

Revision History:

XX/02/1999 - KD

Creation

13/04/2010 - KD

added to SPRAAK