From ace5aeb0f59fdceb99e607af9780added20da37c Mon Sep 17 00:00:00 2001
From: Joseph Redmon <pjreddie@gmail.com>
Date: Fri, 24 Jan 2014 22:51:17 +0000
Subject: [PATCH] MNIST connected network showing off matrices
---
src/connected_layer.c | 146 ++++++++++++++++++++++++++++++------------------
1 files changed, 92 insertions(+), 54 deletions(-)
diff --git a/src/connected_layer.c b/src/connected_layer.c
index 9fafc38..6871b2e 100644
--- a/src/connected_layer.c
+++ b/src/connected_layer.c
@@ -1,96 +1,134 @@
#include "connected_layer.h"
+#include "utils.h"
+#include "mini_blas.h"
#include <math.h>
+#include <stdio.h>
#include <stdlib.h>
#include <string.h>
-connected_layer *make_connected_layer(int inputs, int outputs, ACTIVATION activator)
+connected_layer *make_connected_layer(int inputs, int outputs, ACTIVATION activation)
{
+ fprintf(stderr, "Connected Layer: %d inputs, %d outputs\n", inputs, outputs);
int i;
connected_layer *layer = calloc(1, sizeof(connected_layer));
layer->inputs = inputs;
layer->outputs = outputs;
layer->output = calloc(outputs, sizeof(double*));
+ layer->delta = calloc(outputs, sizeof(double*));
layer->weight_updates = calloc(inputs*outputs, sizeof(double));
+ layer->weight_momentum = calloc(inputs*outputs, sizeof(double));
layer->weights = calloc(inputs*outputs, sizeof(double));
+ double scale = 2./inputs;
for(i = 0; i < inputs*outputs; ++i)
- layer->weights[i] = .5 - (double)rand()/RAND_MAX;
+ layer->weights[i] = rand_normal()*scale;
layer->bias_updates = calloc(outputs, sizeof(double));
+ layer->bias_momentum = calloc(outputs, sizeof(double));
layer->biases = calloc(outputs, sizeof(double));
for(i = 0; i < outputs; ++i)
- layer->biases[i] = (double)rand()/RAND_MAX;
+ //layer->biases[i] = rand_normal()*scale + scale;
+ layer->biases[i] = 0;
- if(activator == SIGMOID){
- layer->activation = sigmoid_activation;
- layer->gradient = sigmoid_gradient;
- }else if(activator == RELU){
- layer->activation = relu_activation;
- layer->gradient = relu_gradient;
- }else if(activator == IDENTITY){
- layer->activation = identity_activation;
- layer->gradient = identity_gradient;
- }
-
+ layer->activation = activation;
return layer;
}
-void run_connected_layer(double *input, connected_layer layer)
+void update_connected_layer(connected_layer layer, double step, double momentum, double decay)
{
- int i, j;
+ int i;
for(i = 0; i < layer.outputs; ++i){
- layer.output[i] = layer.biases[i];
- for(j = 0; j < layer.inputs; ++j){
- layer.output[i] += input[j]*layer.weights[i*layer.inputs + j];
- }
- layer.output[i] = layer.activation(layer.output[i]);
+ layer.bias_momentum[i] = step*(layer.bias_updates[i]) + momentum*layer.bias_momentum[i];
+ layer.biases[i] += layer.bias_momentum[i];
}
-}
-
-void learn_connected_layer(double *input, connected_layer layer)
-{
- calculate_update_connected_layer(input, layer);
- backpropagate_connected_layer(input, layer);
-}
-
-void update_connected_layer(connected_layer layer, double step)
-{
- int i,j;
- for(i = 0; i < layer.outputs; ++i){
- layer.biases[i] += step*layer.bias_updates[i];
- for(j = 0; j < layer.inputs; ++j){
- int index = i*layer.inputs+j;
- layer.weights[index] += step*layer.weight_updates[index];
- }
+ for(i = 0; i < layer.outputs*layer.inputs; ++i){
+ layer.weight_momentum[i] = step*(layer.weight_updates[i] - decay*layer.weights[i]) + momentum*layer.weight_momentum[i];
+ layer.weights[i] += layer.weight_momentum[i];
}
memset(layer.bias_updates, 0, layer.outputs*sizeof(double));
memset(layer.weight_updates, 0, layer.outputs*layer.inputs*sizeof(double));
}
-void calculate_update_connected_layer(double *input, connected_layer layer)
+void forward_connected_layer(connected_layer layer, double *input)
{
- int i, j;
+ int i;
+ memcpy(layer.output, layer.biases, layer.outputs*sizeof(double));
+ int m = 1;
+ int k = layer.inputs;
+ int n = layer.outputs;
+ double *a = input;
+ double *b = layer.weights;
+ double *c = layer.output;
+ gemm(0,0,m,n,k,1,a,k,b,n,1,c,n);
for(i = 0; i < layer.outputs; ++i){
- layer.bias_updates[i] += layer.output[i];
- for(j = 0; j < layer.inputs; ++j){
- layer.weight_updates[i*layer.inputs + j] += layer.output[i]*input[j];
- }
+ layer.output[i] = activate(layer.output[i], layer.activation);
}
}
-void backpropagate_connected_layer(double *input, connected_layer layer)
+void learn_connected_layer(connected_layer layer, double *input)
{
- int i, j;
-
- for(j = 0; j < layer.inputs; ++j){
- double grad = layer.gradient(input[j]);
- input[j] = 0;
- for(i = 0; i < layer.outputs; ++i){
- input[j] += layer.output[i]*layer.weights[i*layer.inputs + j];
- }
- input[j] *= grad;
+ int i;
+ for(i = 0; i < layer.outputs; ++i){
+ layer.delta[i] *= gradient(layer.output[i], layer.activation);
+ layer.bias_updates[i] += layer.delta[i];
}
+ int m = layer.inputs;
+ int k = 1;
+ int n = layer.outputs;
+ double *a = input;
+ double *b = layer.delta;
+ double *c = layer.weight_updates;
+ gemm(0,0,m,n,k,1,a,k,b,n,1,c,n);
}
+void backward_connected_layer(connected_layer layer, double *input, double *delta)
+{
+ memset(delta, 0, layer.inputs*sizeof(double));
+
+ int m = layer.inputs;
+ int k = layer.outputs;
+ int n = 1;
+
+ double *a = layer.weights;
+ double *b = layer.delta;
+ double *c = delta;
+
+ gemm(0,0,m,n,k,1,a,k,b,n,1,c,n);
+}
+/*
+ void forward_connected_layer(connected_layer layer, double *input)
+ {
+ int i, j;
+ for(i = 0; i < layer.outputs; ++i){
+ layer.output[i] = layer.biases[i];
+ for(j = 0; j < layer.inputs; ++j){
+ layer.output[i] += input[j]*layer.weights[i*layer.inputs + j];
+ }
+ layer.output[i] = activate(layer.output[i], layer.activation);
+ }
+ }
+ void learn_connected_layer(connected_layer layer, double *input)
+ {
+ int i, j;
+ for(i = 0; i < layer.outputs; ++i){
+ layer.delta[i] *= gradient(layer.output[i], layer.activation);
+ layer.bias_updates[i] += layer.delta[i];
+ for(j = 0; j < layer.inputs; ++j){
+ layer.weight_updates[i*layer.inputs + j] += layer.delta[i]*input[j];
+ }
+ }
+ }
+ void backward_connected_layer(connected_layer layer, double *input, double *delta)
+ {
+ int i, j;
+
+ for(j = 0; j < layer.inputs; ++j){
+ delta[j] = 0;
+ for(i = 0; i < layer.outputs; ++i){
+ delta[j] += layer.delta[i]*layer.weights[i*layer.inputs + j];
+ }
+ }
+ }
+ */
--
Gitblit v1.10.0